Archive for the english category

Joni

Polling social media users to predict election outcomes

english

The 45th President of the USA

Introduction

The problem of predicting election outcomes with social media is that the data, such as likes, are aggregate, whereas the election system is not — apart from simple majority voting, in which you only have the classic representativeness problem that Gallup solved in 1936. To solve the aggregation problem, one needs to segment the polling data so that it 1) corresponds to the prevailing election system and 2) accurately reflects the voters according to that system. For example, in the US presidential election each state has a certain number of electoral votes. To win, a candidate needs to reach 270 electoral votes.

Disaggregating the data

One obvious solution would be track like sources to profiles and determine the state based on publically given information by the user. This way we could filter out foreign likers as well. However, there are some issues of using likes as indicators of votes. Most importantly, “liking” something on social media does not in itself predict future behavior of an individual to a sufficient degree.

Therefore, I suggest here a simple polling method via social media advertising (Facebook Ads) and online surveys (Survey Monkey). Polling is partly facing the same aforementioned problem of future behavior than using likes as the overarching indicator which is why in the latter part of this article I discuss how these approaches could be combined.

At this point, it is important to acknowledge that online polling does have significant advantages relating to 1) anonymity, 2) cost, and 3) speed. That is, people may feel more at ease expressing their true sentiment to a machine than another human being. Second, the method has the potential to collect a sizeable sample in a more cost-effective fashion than calling. Finally, a major advantage is that due to the scalable nature of online data collection, the predictions can be updated faster than via call-based polling. This is particularly important because election cycles can involve quick and hectic turns. If the polling delay is from a few days to a week, it is too late to react to final week events of a campaign which may still carry a great weight in the outcome. In other words: the fresher the data, the better. (An added bonus is that by doing several samples, we could consider momentum i.e. growth speed of a candidate’s popularity into our model – albeit this can be achieved with traditional polling as well.)

Social media polling (SMP)

The method, social media polling or SMP, is described in the following picture.

Figure 1 Social media polling

The process:

1. Define segmentation criteria

First, we understand the election system. For example, in the US system every state has a certain weight expressed by its share of total electoral votes. There are 50 states, so these become our segmentation criteria. In case we deem appropriate to do further segmentation (e.g., gender, age), we can do so by creating additional segments which are reflected in target groups and surveys. (These sub-segments can also be analyzed in the actual data later on.)

2. Create unique surveys

Then, we create a unique survey for each segment so that the answers will be bucketed. The questions of the survey are identical – they are just behind different links to enable easy segmentation. We create a survey rather than use a visible poll (app) or picture-type of poll (“like if you vote Trump, heart if you vote Hillary”), because we want to avoid social desirability bias. A click on Facebook will lead the user to the unique survey of their segment, and their answers won’t be visible to the public.

3. Determine sample size

Calculating sample size is one of those things that will make your head spin, because there’s no easy answer as to what is a good sample size. Instead, “it depends.” However, we can use some heuristical rules to come up with decent alternatives in the context of elections. Consider two potential sample sizes.

  • Sample size: 500
  • Confidence level: 95%
  • Margin of error: +/- 4.4%
  • Sample size: 1,000
  • Confidence level: 95%
  • Margin of error: +/- 3%

These are seen as decent options among election pollsters. However, the margin of error is still quite sizeable in both of them. For example, if there are two candidates and their “true” support values are A=49%, B=51%, the large margin of error makes us easily go wrong. We could solve this by increasing the sample size, but the problem is that if we would like to reduce the margin of error from +/- 3% to say 1%, our required sample size grows dramatically (more precisely, with a 95% confidence and population size of 1M, it’s 9512 – unpractically high for a 50-state model). In other words, we have to accept the risk of wrong predictions in this type of situation.

All states have over 1,000,000 million people so each of them are considered as “large” populations (this is a mathematical thing – required sample size stabilizes after reaching a certain population size). Although US is characterized as one population, in the context of election prediction it’s actually several different populations (because we have independent states that vote). The procedure we apply is stratified random sampling in which the large general population is split into sub-groups. In practice, each sub-group requires its own sample, and therefore our approach requires a considerably larger sample size than a prediction that would only consider the whole population of the country. But, exactly because of this it should be more accurate.

So, with this lengthy explanation let us say we satisfice with a sample size of 500 per state. That would be 500×50=25,000 respondents. If it would cost 0.60$ to get a respondent via running Facebook ads, the cost for data collection would be 15,000$. For repetitive purposes, there are a few strategies. First, the sample size can be reduced for states that show a large difference between the candidates. In other words, we don’t need to collect a large number of respondents if we “know” the popularity difference between candidates is high. The important thing is that the situation is measured periodically, and sample sizes are flexibly adjusted according to known results. In a similar vein, we can increase the sample size for states where the competition is tight, to reduce the margin of error and therefore to increase the accuracy of our prediction. To my understanding, the opportunity of flexible sampling is not efficiently used by all pollsters.

4. Create Facebook campaigns

For each segment, a target group is created in Facebook Ads. The target group is used to advertise to that particular group; for example, the Michigan survey link is only shown to people from Michigan. That way, we minimize the risk of people outside the segment responding (however, they can excluded later on by IP). At this stage, creating attractive ads help keeping the cost per response low.

5. Run until sample size is reached

The administrator observes the results and stops the data collection once a segment has reached the desired sample size. When all segments are ready, the data collection is stopped.

6. Verify data

Based on IP, we can filter out respondent who do not belong to our particular geographical segment (=state).

Ad clicks can be used to determine sample representativeness by other factors – in other words, we can use Facebook’s campaign reports to segment age and gender information. If a particular group is under-represented, we can correct by altering the targeting towards them and resume data collection. However, we can also accept the under-representation if we have no valid reference model as to voting behavior of the sub-segments. For example, millennials might be under-represented in our data, but this might correspond with their general voting behavior as well – if we assume survey response rate corresponds with voting rate of the segments, then there is no problem.

7. Analyze results

The analysis process is straight-forward:

segment-level results x weights = prediction outcome

For example, in the US presidential election, segment-level results would be each state (who polls highest in the state is the winner there) which would be multiplied by the share of electoral votes of each state. The candidate who gets at least 270 votes is the predicted winner.

Other methods

Now, as for other methods, we can use behavioral data. I have previously argued behavioral data is a stronger indicator of future actions since it’s free from reporting bias. In other words, people say they do, but won’t end up doing. This is a very common problem, but in research and daily life.

To correct for that, we consider two approaches here:

1) The volume of likes method, which parallels a like to a vote (the more likes a candidate has in relation to another candidate, the more likely they are to win)

For this method to work, the “intensity of like”, i.e. its correlation to behavior should be determined, as not all likes are indicators of voting before. Likes don’t readily translate into votes, and there does not appear to be other information we can use to further examine their correlation (like is a like). We could, however, add contextual information of the person, or use rules such as “the more likes a person gives, the more likely (s)he is to vote for a candidate.”

Or, we could use another solution which I think is better:

2) Text analysis/mining

By analyzing social media comments of a person, we can better infer the intensity of their attitude towards a given topic (in this case, a candidate). If a person is using strongly positive vocabulary while referring to a candidate, then (s)he is more likely to vote for him/her than if the comments are negative or neutral. Notice that the mere positive-negative range is not enough, because positivity has degrees of intensity we have to consider. It is different to say “he is okay” than “omg he is god emperor”. The more excitement and associated feelings – which need to be carefully mapped and defined in the lexicon – a person exhibits, the more likely voting behavior is.

Limitations

As I mentioned, even this approach risks shortcoming of representativeness. First, the population on Facebook may not correspond with the population at large. It may be that the user base is skewed by age or some other factor. The choice of platform greatly influences the sample; for example, SnapChat users are on average younger than Facebook users, whereas Twitter users are more liberal. It is not clear whether Facebook’s user base represents a skewed sample or not. Second, the people voicing their opinions may be part of “vocal minority” as opposed to “silent majority”. In that case, we apply the logic of Gaussian standard distribution and assumed that the general population is more lenient to middle ground than the extremes — if, in addition, we would assume the central tendency to be symmetrical (meaning people in the middle are equally likely to tip into either candidate in a dual race), the analysis of extremes can still yield a valid prediction.

Another limitation may be that advertising targeting is not equivalent to random sampling, but has some kind of bias. That bias could emerge e.g. from 1) ad algorithm favoring a particular sub-set of the target group, i.e. showing more ads to them, whereas we would like to get all types of respondents; or 2) self-selection in which the respondents are of similar kind and again not representative to the population. Out of my head, I’d say number two is less of a problem because those people who show enough interest are also the ones who vote – remember, essentially we don’t need to care about the opinions of the people who don’t vote (that’s how elections work!). But number one could be a serious issue, because ad algorithm directs impressions based on responses and might identify some hidden pattern we have no control over. Basically, the only thing we can do is examine superficial segment information on the ad reports, and evaluate if the ad rotation was sufficient or not.

Combining different approaches

As both approaches – traditional polling and social media analysis – have their shortcomings and advantages, it might be feasible to combine the data under a mixed model which would factor in 1) count of likes, 2) count of comments with high affinity (=positive sentiment), and 3) polled preference data. A deduplicating process would be needed to not count twice those who liked and commented – this requires associating likes and comments to individuals. Note that the hybrid approach requires geographic information as well, because otherwise segmentation is diluted. Anyhow, taking user as the central entity could be a step towards determining voting propensity:

user (location, count of likes, count of comments, comment sentiment) –> voting propensity

Another way to see this is that enriching likes with relevant information (in regards to the election system) can help model social media data in a more granular and meaningful way.

Joni

Analyzing sentiment of topical dimensions in social media

english

Introduction

Had an interesting chat with Sami Kuusela from Underhood.co. Based on that, got some inspiration for an analysis framework which I’ll briefly describe here.

The model

Figure 1 Identifying and analyzing topical text material

The description

  1. User is interested in a given topic (e.g., Saara Aalto, or #saaraaalto). He enters the relevant keywords.
  2. The system runs a search and retrieves text data based on that (e.g., tweets).
  3. A cluster analysis (e.g., unsupervised topic model) identifies central themes from the data.
  4. Vectorization of representative keywords based on cluster analysis (e.g., 10 most popular) is run to extract words from a reference lexicon of words that have a similar meaning. This increases the generality of each topic cluster by associating them with other words that are close in the vector space.
  5. Text mining is run to refine the themes, i.e. placing the right text pieces under the correct themes. These are now called “dimensions”, since they describe the key dimensions of the text corpus (e.g., Saara’s voice, performance, song choices…).
  6. Sentiment analysis can be run to score the general (pos/neg/neu) or specific (e.g., emotions: joy, excitement, anger, disappointment, etc.) sentiment of each dimension. This could be done by using a machine-learning model with annotated training data (if the data-set is vast), or some sentiment lexicon (if the data-set is small).

I’m not sure whether steps 4 and 5 would improve the system’s ability to identify topics. It might be that a more general model is not required because the system already can detect the key themes. Would be interesting to test this with a developer.

Anyway, what’s the whole point?

The whole point is to acknowledge that each large topic naturally divides into small sub-topics, which are dimensions that people perceive relevant for that particular topic. For example, in politics it could be things like “economy”, “domestic policy”, “immigration”, “foreign policy”, etc. While the dimensions can have some consistency based on the field, e.g. all political candidates share some dimensions, the exact mix is likely to be unique, e.g. dimensions of social media texts relating to Trump are likely to be considerably different from those of Clinton. That’s why the analysis ultimately needs to be done case-by-case.

In any case, it is important to note that instead of giving a general sentiment or engagement score of, say a political candidate, we can use an approach like this to give a more in-depth or segmented view of them. This leads to better understanding of “what works or not”, which is information that can be used in strategic decision-making. In addition, the topic-segmented sentiment data could be associated with predictors in a predictive model, e.g. by multiplying each topic sentiment with the weight of the respective topic (assuming the topic corresponds with the predictor).

Limitations

This is just a conceptual model. As said, would be interesting to test it. There are many potential issues, such as handling with cluster overlap (some text pieces can naturally be placed into several clusters which can cause classification problems) and hierarchical issues (e.g., “employment” is under “economy” and should hence influence the former’s sentiment score).

Joni

Agile methods for predicting contest outcomes by social media analysis

english

People think, or seem to assume, that there is some magical machine that spits out accurate predictions of future events from social media data. There is not, and that’s why each credible analysis takes human time and effort. But therein also lies the challenge: when fast decisions are needed, time-taking analyses reduce agility. Real-time events would require real-time analysis, whereas data analysis is often cumbersome and time-taking effort, including data collection, cleaning, machine training, etc.

It’s a project for weeks or days, not for hours. All the practical issues of the analysis workflow make it difficult to provide accurate predictions at a fast pace (although there are other challenges as well).

An example is Underhood.co – they predicted Saara Aalto to win X-Factor UK based on social media sentiment, but ended up being wrong. While there are many potential reasons for this, my conclusion is that their indicators lack sufficient predictive power. They are too reliant on aggregates (in this case country-level data), and had a problematic approach to begin with – just like with any prediction, the odds change on the go as new information becomes available, so you should never predict the winner weeks ahead. Of course, theirs was just a publicity stunt where they hoped being right would prove the value of their service. Another example, of course, is the US election where prediction markets were completely wrong of the outcome. That was, according to my theory, because of wrong predictors – polls ask what is your preference or what you would do, whereas social media engagement shows what people do (in social media), and as such are closer to real behavior, hence better predictors.

Even if I do think human analysts are still needed in the near future, more solutions for quick collection and analysis of social media data are needed, especially to combine the human and machine work in the best possible way. Some of these approaches can be based on automation, but others can be methodological, such as quick definition of relevant social media outlets for sampling.

Here are some ideas I have been thinking of:

I. Data collection

  1. Quick definition of choice space (e.g., candidates in a political election, X-Factor contestants)
  2. Identification of related media social media outlets (i.e., communities, topic hashtags)
  3. Collecting sample (API, scraping, or copy-paste (crowdsourcing))

Each part is case-dependent and idiosyncratic – for whatever event, I’m thinking competitions here, you have to this work from scratch. Ultimately, you cannot get the whole Internet as your data, but you want the sample to be as representative as possible. For example, it was obvious that Twitter users showed much more negative sentiment towards Trump than Facebook users, and in both platforms you had supporter groups/topic concentrations that should first be identified before any data collection. Then, the actual data collection is tricky. People again seem to assume all data is easily accessible. It’s not – while Twitter and Facebook have an API, Youtube and Reddit don’t, for example. This means the comments that you use for predicting the outcome (by analyzing their relative share of the total as well as the strength of the sentiment beyond pos/neg) need to be fetched either by webscraping or manually copying them to a spreadsheet. Due to large volumes of data, crowdsourcing could be useful — e.g., setting up a Google Sheet where crowdworkers each paste the text material in clean format. The raw text content, e.g. tweets, Facebook comments, Reddit comments, is put in separate sheets for each candidate.

II. Data analysis

  1. Cluster visualization (defining clusters, visualizing their respective sizes (plot # of voters), breakdown by source platform and potential other factors)
  2. Manual training (classifying the sentiment, or “likelihood to vote”)
  3. Machine classification (calculating the number of likely voters)

In every statistical analysis, the starting point should be visualizing the data. This shows an aggregate “helicopter view” of the situation. Such a snapshot is useful also for demonstrating the results for the end-user, to let the data speak for itself. Candidates are bubbles in the chart, their sizes in respect to the number of calculated likely voters. The data could be broken down according to source platforms, or other factors, by using the candidate as a point of gravity for the cluster.

Likelihood to vote could be classified as a scale, not binary. That is, instead of saying “sentiment is positive: YES/NO”, we could say “How likely is the person to vote?” which is the same as asking how enthusiastic or engaged he or she is. Therefore, a scale is better, e.g. ranging from -5 (definitely not voting for this candidate) to +5 (definitely voting for this candidate). The manual training, which also could be done with the help of crowd, helps the machine classifier to improve its accuracy on the go. Based on training data, it would generalize classification to all material. Now, the material is bucketed so that each candidate is evaluated separately and the number of likely voters can be calculated. It is possible that the machine classifier could benefit from training input from both candidates, inasmuch the language showing positive and negative engagement is not significantly different.

It is important to note that negative sentiment does not really matter. What we are interested in is the number of likely voters. This is because of the election dynamics – it does not matter how poor a candidates aggregate sentiment is, i.e. the ratio between haters and sympathizers, as long as his or her number of likely voters is higher than that of the competition. This effect was evident in the recent US presidential election.

The crucial thing is keep the process alive during the whole election/competition period. There is no such point where it becomes certain that one loses and the other remains, although the divide can become substantial and therefore increase the accuracy of the prediction.

III. Presentation of results

  • constantly updating feed (à la Facebook video stream)
  • cluster visualization
  • search trend widget (source: Google Trends)
  • live updating predictions (manual training –> machine model)

The results could be shown as a form of dashboard to the end user. Search trend graph and the above mentioned cluster visualization could be viable alternatives. In addition, it would be interesting to see the count of voters evolving in time – in such a way that it, along with the visualization, could be “played back” to examine the development in time. In other words, interactive visualization. As noted, the prediction, or the count of likely votes, should update real-time as a result of combined human-machine work.

Conclusion and discussion

The idea behind development of more agile methods to use social media data to predict content outcomes is that the accuracy of the prediction is based on the choice of indicators rather than the finesse of the method. For example, complex Bayesian models falsely predicted Hillary Clinton would win the election. It’s not that the models were poorly built; they just used the wrong indicators, namely polling data. This is the usual case of ‘garbage in, garbage out’, and it shows that the choice of indicators is more important than technical features of the predictive model.

The choice of indicators should be done based on their predictive power and although I don’t have strict evidence on it, it intuitively makes sense that social media engagement is a stronger indicator in many instances than survey data, because it’s based on actual preferences instead of stated preferences. Social scientists know from long tradition of survey research that there are myriad of social effects reducing the reliability of data (e.g., social desirability bias). Those, I would argue, are much smaller issue in social media engagement data.

However, to be fair, there can be issues of bias in the social media engagement data. The major concern is low participation rate: a common heuristic is that 1/10 of participants actually contribute in writing, while the other 9/10 are readers whose real thoughts remain unknown. It’s then a question of how well does the vocal minority reflect the opinion of the silent majority. Or, in some cases, this is irrelevant for competitions if the overall voting share remains low. For example, if it’s 60% it is relative much more important to mobilize the active base than if voting was close to 100% where one would need a universal acceptance.

Another issue is the non-representative sampling. This is a concern when the voting takes place offline, and online data does not accurately reflect the voting of those who do not express themselves online. However, as social media participation is constantly increasing, this becomes less of a problem. In addition, compared to other methods of data collection – apart from stratified polling, perhaps – social media is likely to give a good result on competitive predictions because of their political nature. People who strongly support a candidate are more likely to be vocal about it, and the channel for voicing their opinion is the social media.

It is evident that the value of social media engagement as a predictor is currently underestimated, as proven by the large emphasis put on political polls and virtually zero discussion on social media data. As a direct consequence of this, those who are able to leverage the social media data in the proper way will gain competitive advantage, be it betting market, or any other purpose where prediction accuracy plays a key role. The prediction work will remain a hybrid effort by man and machine.

Joni

On complexity of explaining business failure

english
On complexity of explaining business failure

Introduction

During the research period for my dissertation based on startup failures, I realized there are multiple layers of failure factors associated with any given company (or, in reverse, success factors).

These are:

  1. generic business problems (e.g., cash-flow)
  2. individual-level problems (e.g., personal chemistry)
  3. company type problems (e.g., lack of funding for startups)
  4. business model problems (e.g., chicken-and-egg for platforms)

Only if you combine these multiple layers – or perspectives – can you understand why one business venture fails and another one succeeds. However, it is also a relative and interpretative task — I would argue there can be no objective dominant explanation but failure as an outcome is always a combination of reasons and cannot therefore be reduced into simple explanations at all.

A part of the reason for the complexity is the existence of parallel root causes.

For example,

  • A company can said to have failed because it runs out of money.
  • However, why did it run out of money? Because customers would not buy.
  • Why didn’t they buy? Because the product was bad.
  • Why was the product bad? Because the team failed to recognize true need in the market.
  • Why did they fail to recognize it? They lacked such competence.
  • Why did they lack the competence? Because they had not enough funding to acquire it.

Alas! We ended up making a circular argument. That can happen with any failure explanation, as can coming up with a different root cause. In a team of many, while also considering several stakeholders, it is common that people’s explanations to cause and effect vary a great deal. It is just a feature of social reality that we have a hard time of finding unambiguity.

Conclusion

In general, it is hard to dissect cause and effect. Human beings are inclined to form narratives where they choose a dominant explanation and discard others. By acknowledging a multi-layered view on failure, one can examine a business case by applying different lenses one after another. This includes interviewing different stakeholder groups and understanding multiple perspectives ranging from individual to structural issues.

There are no easy answers as to why a particular company succeeds or fails, even though the human mind and various success stories would lead you to believe so!

Joni

Buying? How to determine the offer price for a website

english

Introduction

A few years back I was considering of buying a website. In the end, I didn’t end up making the offer, largely because I couldn’t figure out how to calculate the offer price in a plausible way. Since then I’ve had a bit more experience in estimating figures in other contexts, as well as participating in some M&A discussions in the ecommerce field. But today, while cleaning my inbox, I happened to read that old email from many years ago, and thought of sharing some thoughts on the topic — hopefully as a bit wiser person!

Basic figures

If you are planning of buying a website and thinking about the offer price, you should know some basic figures of the website:

ARPU, or average revenue per user

if there is none, you have to estimate the earning potential. If the monetization model is advertising, find some stats about avg. CPMs in the industry. If it’s freemium, consider avg. revenue per premium user as well as the conversion rate from free to paid (again, you can find some industry averages).

Number of users/visitors

This is easy to get from analytics software.

Revenue

Revenue or revenue potential (if there is none at the moment) can be calculated by multiplying the two previous figures. So you would move from unit metrics to aggregate numbers.

Profit

You also need to consider the cost of maintenance, marketing and other actions that are needed to keep the site running and growing. Deduct those from the revenue to get profit. If you want faster growth, you need to factor in an investment for that; although it’s not exactly a part of the offer calculation, it still needs to be considered in the overall plan for making money with the website.

Calculating the offer price

Then, to determine offer price you need to multiply the profit with a time unit, e.g. months or years, to get the offer price. This figure is like a line in the sand — you can try and think it from the seller’s perspective: how many years or months of profit would he want to recoup in order for him to be willing to sell.

As an investor, your best break can be found when the profit is low, but revenue potential and number of visitors as well as visitor loyalty are high. The high revenue potential means that there is likely to be a realistic monetization model, but because that has not been applied yet, one can negotiate a good price if the seller is willing to let go of the website. Loyalty – manifested in high rate of returning visitors – indicates that the website provides real value for its visitors instead of relying e.g. on spammy tactics to lure in casual browsers. In the end, the quality of traffic matters a lot in whatever business model you apply.

You should also consider the stability of the figures – in particular, the historical growth rate. With the historical growth rate, you are able to project the development of traffic and revenue in the future. At this point, be realistic of what it takes to uphold the growth rate and thorough in asking the current owner in great detail what he has done so far and why. This information is highly valuable.

Because there is a lot of imprecision in coming up with the aforementioned figures, you would be wise to factor in risk at every stage of the calculation. Convey the risk also to the buyer in a credible way, so that he sees ‘it won’t be easy’ to get your money back. This is a negotiation tactic but also the real state of affairs in many cases.

Closing remarks

I don’t include any “goodwill” on things like brand or design in the calculation, because I think those are irrelevant for the price determination. All sunk costs that don’t serve the revenue potential are pretty much redundant — sticking to real numbers and, when they are absent, realistic estimates — is a much better way of determining the price of a website.

Joni

How to measure media bias?

english

Mass media is old, and so is their bias.

Introduction

Media bias is under heavy discussion at the moment, especially relating to the on-going presidential election in the US. However, the quality of discussion is not the way it should be; I mean, there should be objective analysis on the role of the media. Instead, most comments are politically motivated accusations or denials. This article aims to be objective, discussing the measurement of media bias; that is, how could we identify whether a particular media outlet is biased or not? The author feels there are not generally acknowledged measures for this, so it is easy to claim or deny bias without factual validation. Essentially, this erodes the quality of the discussion, leading only into a war of opinions. Second, without the existence of such measures, both the media and the general public are unable to monitor the fairness of coverage.

Why is media fairness important?

Fairness of the media is important for one main reason: the media have a strong influence on the public opinion. In other words, journalists have great power, and with great power comes great responsibility. The existence of bias leads to different standards of coverage depending on the topic being reported. In other words, the information is being used to portray a selective view of the world. This is analogous to confirmation bias; a person wants to prove a certain point, so he or she only acknowledges evidence supporting that point. Such behavior is very easy for human beings, for which reason journalists should be extra cautious in letting their own opinions influence the content of their reportage.

In addition to being a private problem, the media bias can also be understood as a systemic problem. This arises through 1) official guidelines and 2) informal group think. First, the official guidelines means that the opinions, beliefs or worldviews of the particular media outlet are diffused down the organization. Meaning that the editorial board communicates its official stance (“we, as a media outlet, support a political candidate X”) which is then taken by the individual reporters as their ethos. When the media outlet itself, or the surrounding “media industry” as a whole, absorbs a view, there is a tendency to silence the dissidents. This, again, can be reduced to elementary human psychology, known as the conformity bias or group think. Because others in your reference group accept a certain viewpoint, you are more likely to accept it as well due to social pressure. The informal dynamics are even more dangerous to objective reporting than the official guidelines because they are subtle and implicit by nature. In other words, journalist may not be aware of bias and just consider their worldview “normal” while arguments opposing it are classified as wrong and harmful.

Finally, media fairness is important due to its larger implications on information sources and the actions taken by citizens based on the information they are exposed to. It is in the society’s best interest that people resort to legitimate and trustworthy sources of information, as opposed to unofficial, rogue sources that can spread misinformation or disinformation. However, when the media becomes biased, it loses its legitimacy and becomes discredited; as a form of reactance to the biased stories, citizens turn to alternative sources of information. The problem is that these sources may not be trustworthy at all. Therefore, by waving their journalistic ethics, the mass media become at par with all other information sources; in a word, lose their credibility. The lack of credible sources of information leads into a myriad of problems for the society, such as distrust in the government, civil unrest or other forms of action people take based on the information they receive. Under such circumstances, the problem of “echo chamber” is fortified — individuals feel free to select their sources according to their own beliefs instead of facts. After all, if all information is biased, what does it matter which one you choose to believe in?

How to measure media bias?

Overview

While it may not be difficult to define media bias at a general level, it may be difficult to observe an instance of bias in an unanimously acceptable way. That is where commonly accepted measures could be of some help. To come up with such measures, we can start by defining the information elements that can be retrieved for objectivity analysis. Then, we should consider how they can best be analyzed to determine whether a particular media outlet is biased.

In other words, what information do we have? Well, we can observe two sources: 1) the media itself, and 2) all other empirical observations (e.g., events taking place). Notice that observing the world only through media would be inaccurate testimony of human behavior; we draw a lot from our own experiences and from around us. By observing the stories created by the media we know what is being reported and what is not being reported. By observing things around us (apart from the media), we know what is happening and what is not happening. By combining these dimensions, we can derive

  1. what is being reported (and happens)
  2. what is being reported (but does not happen)
  3. what is not being reported (but happens), and
  4. what is not being reported (but does not happen).

Numbers 2 and 4 are not deemed relevant for this inquiry, but 1 and 3 are. Namely, the choice of information, i.e. what is being reported and what is being left out of reporting. Hence, this is the first dimension of our measurement framework.

1. Choice of information

  • topic inclusion — what topics are reported (themes –> identify, classify, count)
  • topic exclusion — what topics are not reported (reference –> define, classify, count)
  • story inclusion — what is included in the reportage (themes –> identify, classify, count)
  • story exclusion — what is left out of the reportage (reference –> define, classify, count)
  • story frequency — how many times a story is repeated (count)

This dimension measures what is being talked about in the media. It measures inclusion, exclusion and frequency to determine what information the media disseminates. The two levels are topics and stories — both have themes that can be identified, then material classified into them, and counted to get an understanding of the coverage. Measuring exclusion works in the same way, except the analyst needs to have a frame of reference he or she can compare the found themes with. For example, if the frame of reference contains “Education” and the topics found from the material do not include education, then it can be concluded that the media at the period of sampling did not cover education. Besides themes, reference can include polarity, and thus one can examine if opposing views are given equal coverage. Finally, the frequency of stories measures media’s emphasis; reflecting the choice of information.

Because all information is selected from a close-to-infinite pool of potential stories, one could argue that all reportage is inherently biased. Indeed, there may not be universal criteria that would justify reporting Topic A over Topic B. However, measurement helps form a clearer picture of a) what the media as a whole is reporting, and b) what does each individual media outlet report in comparison to others. A member of the audience is then better informed on what themes the media has chosen to report. This type of helicopter view can enhance the ability to detect a biased information choice, either by a particular media outlet or the media as a whole.

The question of information choice is pertinent to media bias, especially relating to exclusion of information. A biased reporter can defend himself by arguing “If I’m biased, show me where!”. But bias is not the same as inaccuracy. A biased story can still be accurate, for example, it may only leave some critical information out. The emphasis of a certain piece of information at the expense of other is a clear form of bias. Because not every piece of information can be included in a story, something is forcefully let out. Therefore, there is a temptation to favor a certain story-line. However, this concern can be neutralized by introducing balance; for a given topic, let there be an equal effort for exhibiting positive and negative evidence. And in terms of exclusion, discarding an equal amount of information from both extremes, if need be.

In addition to measuring what is being reported, we also need to consider how it is being reported. This is the second dimension of the measurement framework, dealing with the formulation of information.

2. Formulation of information

  • IN INTERVIEWS: question formulation — are the questions reporters are asking neutral or biased in terms of substance (identify, classify, count)
  • IN REPORTS: message formulation — are the paragraphs/sentences in reportage neutral or biased in terms of substance (classify, count)
  • IN INTERVIEWS: tone — is the tone reporters are asking the questions neutral or biased (classify count)
  • IN REPORTS: tone — are the paragraphs/sentences in reportage neutral or biased in terms of tone (classify, count)
  • loaded headlines (identify, count)
  • loaded vocabulary (identify, count)
  • general sentiment towards key objects (identify, classify: pos/neg/neutral)

This dimension measures how the media reports on the topics it has chosen. It is a form of content analysis, involving qualitative and quantitative features. Measures cover interview type of settings, as well as various reportages such as newspaper articles and television coverage. The content can be broken down into pieces (questions, paragraphs, sentences) and their objectivity evaluated based on both substance and tone. An example of bias in substance would be presenting an opinion as a fact, or taking a piece of information out of context. An example of biased tone would be using negative or positive adjectives in relation to select objects (e.g., presidential candidates).

Presenting loaded headlines and text as percentage of total observations gives an indication of how biased the content is. In addition, the analyst can evaluate the general sentiment the reportage portrays of key objects — this includes first identifying the key objects of the story, and then classifying their treatment on a three-fold scale (positive, negative, neutral).

I mentioned earlier that agreeing on the observation of bias is an issue. This is due to the interpretative nature of these measures; i.e., they involve a degree of subjectivity which is generally not considered as a good characteristic for a measure. Counting frequencies (e.g., how often a word was mentioned) is not susceptible to interpretation but judging the tone of the reporter is. Yet, those are the kind of cues that reveal a bias, so they should be incorporated in the measurement framework. Perhaps we can draw an analogy to any form of research here; it is always up to the integrity of the analyst to draw conclusions. Even studies that are said to include high reliability by design can be reported in a biased way, e.g. by re-framing the original hypotheses. Ultimately, application of measurement in social sciences remains at the shoulder of the researcher. Any well-trained, committed researcher is more likely to follow the guideline of objectivity than not; but of course this cannot be guaranteed. The explication of method application should reveal to an outsider the degree of trustworthiness of the study, although the evaluation requires a degree of sophistication. Finally, using several analysts reduces an individual bias in interpreting content; inter-rater agreement can then be calculated with Cohen’s kappa or similar metrics.

After assessing the objectivity of the content, we turn to the source. Measurement of source credibility is important in both validating prior findings as well as understanding why the (potential) bias takes place.

3. Source credibility

  • individual political views (identify)
  • organizational political affiliation (identify)
  • reputation (sample)

This dimensions measures why the media outlet reports the way it does. If individual and organizational affiliations are not made clear in the reportage, the analyst needs to do work to discover them. In addition, the audience has shaped a perception of bias based on historical exposure to the media outlet — running a properly sampled survey can provide support information for conclusions of the objectivity study.

How to prevent media bias?

The work of journalists is sometimes compared to that of a scientist: in both professions, one needs curiosity, criticality, ability to observe, and objectivity. However, whereas scientists mostly report dull findings, reporters are much more pressured to write sexy, entertaining stories. This leads into the the problem of sense-making, i.e. reporters create a coherent story with a clear message, instead showing the messy reality. The sense-making bias in itself favors media bias, because creating a narrative forces one to be selective of what to include and what to exclude. As long as there is this desire for simple narratives, coverage of complex topics cannot be entirely objective. We may, however, mitigate this effect by upholding certain principles.

I suggest three principles for the media to uphold in their coverage of topics.

  • criticality
  • balance
  • objectivity
  • independence

First, the media should have a critical stance to its object of reportage. Instead of accepting the piece of information they receive as truth, they should push to ask hard questions. But that should be done in a balanced way – for example, in a presidential race, both candidates should get an equal amount of “tough” questions. Furthermore, journalists should not absorb any “truths”, beliefs or presumptions that affect in their treatment of a topic. Since every journalist is a human being, this requirement is quite an idealistic one; but the effect of personal preferences or those imposed by the social environment should in any case be mitigated. The goal of objectivity should be cherished, even if the outcome is in conflict with one’s personal beliefs. Finally, the media should be independent. Both in that it is not being dictated by any interest group, public or private, on what to report, but also in that it is not expressing or committing into a political affiliation. Much like church and state are kept separate according to Locke’s social contract as well as Jefferson’s constitutional ideas, the press and the state should be separated. This rule should apply to both publicly and privately funded media outlets.

Conclusion

The status of the media is precious. They have an enormous power over the opinions of the citizens. However, this is conditional power; should they lose objectivity, they’d also lose the influence, as people turn to alternative sources of information. I have presented that a major root cause of the problem is the media’s inability to detect its own bias. Through better detection and measurement of bias, corrective action can be taken. But since those corrective actions are conditioned to willingness to be objective, a willingness many media outlets are not signalling, the measurement in itself is not adequate in solving the larger problem. At a larger scale, I have proposed there be a separation of media and politics, which prevents by law any media outlet to take a political side. Such legislation is likely to increase objectivity and decrease the harmful polarization that the current partisan-based media environment constantly feeds into.

Overall, there should be some serious discussion on what the role of media in the society should be. In addition, attention to journalistic education and upholding of journalistic ethics should be paid. If the industry is not able to monitor itself, it is upon the society to introduce such regulation that the media will not abuse its power but remains objective. I have suggested the media and related stakeholders provide information on potential bias. I have also suggested new measures for bias that consider both the inclusion and exclusion of information. The measurement of inclusion can be done by analyzing news stories for common keywords and themes. If the analyst has an a prior framework of topics/themes/stories he or she considers as reference, it can be then concluded how well the media covers those themes by classifying the material accordingly. Such analysis would also reveal what is not being reported, an important distinction that is often not taken into account.

Joni

Defining SMQs: Strategic Marketing Questions

english

Introduction

Too often, marketing is thought of being advertising and nothing more. However, already Levitt (1960) and Kotler (1970) established that marketing is a strategic priority. Many organizations, perhaps due to lack of marketers in their executive boards, have since forgotten this imperative.

Another reason for decreased importance of marketing is due to marketing scholars pushing the idea that “everything is marketing” which leads to decay of the marketing concept – if it is everything, it is nothing.

Nevertheless, if we reject the omni-marketing concept and return to the useful way of perceiving marketing, we observe the linkage between marketing and strategy.

Basic questions

Tania Fowler wrote a great piece on marketing, citing some ideas of Professor Roger Martin’s HBR article (2014). Drawing from that article, the basic strategic marketing questions are:

  • Who are our customers? (segmentation)
  • Why do they care about our product? (USPs/value propositions/benefits)
  • How are their needs and desires evolving? (predictive insight)
  • What potential customers exist and why aren’t we reaching them? (market potential)

This is a good start, but we need to expand the list of questions. Borrowing from Osterwalder (2009) and McCarthy (1960), let’s apply BMC (9 dimensions of a business model) and 4P marketing mix thinking (Product, Place, Promotion, Price).

Business Model Canvas approach

This leads to the following set of questions:

  • What is the problem we are solving?
  • What are our current revenue models? (monetization)
  • How good are they from customer perspective? (consumer behavior)
  • What is our current pricing strategy? (Kotler’s pricing strategies)
  • How suitable is our pricing to customers? (compared to perceived value)
  • How profitable is our current pricing?
  • How competitive is our current pricing?
  • How could our pricing be improved?
  • Where are we distributing the product/solution?
  • Is this where customers buy similar products/solutions?
  • What are our potential revenue models?
  • Who are our potential partners? Why? (nature of win-win)

Basically, each question can be presented as a question of “now” and “future”, whereupon we can identify strategic gaps. Strategy is a lot about seeing one step ahead — the thing is, foresight should be based on some kind of realism, or else fallacies take the place of rationality. Another point from marketing and startup literature is that people are not buying products, but solutions (solution-based selling, product-market fit, etc.) Someone said the same thing about brands, but I think solution is more accurate in the strategic context.

Adding competitors and positioning

The major downside of BMC and 4P thinking from strategic perspective is their oversight of competition. Therefore, borrowing from Ries and Trout (1972) and Porter (1980), we add these questions:

  • Who are our direct competitors? (substitutes)
  • Who are our indirect competitors? (cross-verticality, e.g. Google challenging media companies)
  • How are we different from competitors? (value proposition matrix)
  • Do our differentiating factors truly matter to the customers? (reality check)
  • How do we communicate our main benefits to customers? (message)
  • How is our brand positioned in the minds of the customers? (positioning)
  • Are there other products customers need to solve their problem? What are they? (complements)

Defining the competitive advantage, or critical success factors (CSFs), leads into natural linkage to resources, as we need to ask what are the resources we need to execute, and how to acquire and commit those resources (often human capital).

Resource-based view

Therefore, I’m turning to resource-based thinking in asking:

  • What are our current resources?
  • What are the resources we need to be competitive? (VRIN framework)
  • How to we acquire those resources? (recruiting, M&As)
  • How do we commit those resources? (leadership, company culture)

Indeed, company culture is a strategic imperative which is often ignored in strategic decision making. Nowadays, perhaps more than ever, great companies are built on talent and competence. Related strategic management literature deals with dynamic capabilities (e.g., Teece, 2007) and resource-based view (RBV) (e.g., Wernerfelt, 1984). In practice, companies like Facebook and Google do everything possible to attract and retain the brightest minds.

Do not forget profitability

Finally, even the dreaded advertising questions have a strategic nature, relating to customer acquisition and loyalty, as well as ROI in regards to both as well as to our offering. Considering this, we add:

  • How much does it cost to acquire a new customer?
  • What are the best channels to acquire new customers?
  • Given the customer acquisition cost (CAC) and customer lifetime value (CLV), are we profitable?
  • How profitable are each products/product categories? (BCG matrix)
  • How can we make customers repeat purchases? (cross-selling, upselling)
  • What are the best channels to encourage repeat purchase?
  • How do we encourage customer loyalty?

As you can see, these questions are of strategic nature, too, because they are directly linked to revenue and customer. After all, business is about creating customers, as stated by Peter Drucker. However, Drucker also maintained that a business with no repeat customers is no business at all. Thus, marketing often focuses on customer acquisition and loyalty.

The full list of strategic marketing questions

Here are the questions in one list:

  1. Who are our customers? (segmentation)
  2. Why do they care about our product? (USPs/value propositions/benefits)
  3. How are their needs and desires evolving? (predictive insight)
  4. What potential customers exist and why aren’t we reaching them? (market potential)
  5. What is the problem we are solving?
  6. What are our current revenue models? (monetization)
  7. How good are they from customer perspective? (consumer behavior)
  8. What is our current pricing strategy? (Kotler’s pricing strategies)
  9. How suitable is our pricing to customers? (compared to perceived value)
  10. How profitable is our current pricing?
  11. How competitive is our current pricing?
  12. How could our pricing be improved?
  13. Where are we distributing the product/solution?
  14. Is this where customers buy similar products/solutions?
  15. What are our potential revenue models?
  16. Who are our potential partners? Why? (nature of win-win)
  17. Who are our direct competitors? (substitutes)
  18. Who are our indirect competitors? (cross-verticality, e.g. Google challenging media companies)
  19. How are we different from competitors? (value proposition matrix)
  20. Do our differentiating factors truly matter to the customers? (reality check)
  21. How do we communicate our main benefits to customers? (message)
  22. How is our brand positioned in the minds of the customers? (positioning)
  23. Are there other products customers need to solve their problem? What are they? (complements)
  24. What are our current resources?
  25. What are the resources we need to be competitive? (VRIN framework)
  26. How to we acquire those resources? (recruiting, M&As)
  27. How do we commit those resources? (leadership, company culture)
  28. How much does it cost to acquire a new customer?
  29. What are the best channels to acquire new customers?
  30. Given the customer acquisition cost (CAC) and customer lifetime value (CLV), are we profitable?
  31. How profitable are each products/product categories? (BCG matrix)
  32. How can we make customers repeat purchases? (cross-selling, upselling)
  33. What are the best channels to encourage repeat purchase?
  34. How do we encourage customer loyalty?

The list should be universally applicable to all companies. But filling in the list is not “oh, let me guess” type of exercise. As you can see, answering to many questions requires customer and competitor insight that, as the startup guru Steve Blank says, needs to be retrieved by getting out of the building. Those activities are time-consuming and costly. But only if the base information is accurate, strategic planning serves a purpose. So don’t fall prey to guesswork fallacy.

Implementing the list

One of the most important things in strategic planning is iteration — it’s not “set and forget”, but “rinse and repeat”. So, asking these questions should be repeated from time to time. However, people tend to forget repetition. That’s why corporations often use consultants — they need fresh eyes to spot opportunities they’re missing due to organizational myopia.

Moreover, communicating the answers across the organization is crucial. Having a shared vision ensures each atomic decision maker is able to act in the best possible way, enabling adaptive or emergent strategy as opposed to planned strategy (Mintzberg, 1978). For this to truly work, customer insight needs to be internalized by everyone in the organization. In other words, strategic information needs to be made transparent (which it is not, in most organizations).

And for the information to translate into action, the organization should be built to be nimble; empowering people, distributing power and reducing unnecessary hierarchy. People are not stupid: give them a vision and your trust, and they will work for a common cause. Keep them in silos and treat them as sub-ordinates, and they become passive employees instead of psychological owners.

Concluding remarks

We can say that marketing is a strategic priority, or that strategic planning depends on the marketing function. Either way, marketing questions are strategic questions. In fact, strategic management and strategic marketing are highly overlapping concepts. Considering both research and practice, their division can be seen artificial and even counter-productive. For example, strategic management scholars and marketing scholars may speak of the same things with different names. The same applies to the relationship between CEOs and marketing executives. Joining forces reduces redundancy and leads to a better future of strategic decision-making.

Joni

Meaningless marketing

english
Meaningless marketing

I’d say 70% of marketing campaigns have little to no real effect. Most certainly they don’t have a positive return in hard currency.

Yet, most marketers spend their time running around, planning all sorts of campaigns and competitions people couldn’t care less of. They are professional producers of spam, where in fact they should be focusing on core of the business: understanding why customers buy, how could they buy more, what sort of products should we make, how can the business model be improved, etc. The wider concept of marketing deals with navigating the current and the future market; it is not about making people buy stuff they don’t need.

To a great extent, I blame the marketing education. In the academia, we don’t really get the real concept of marketing into our students’ minds. Even the students majoring in marketing don’t truly “get” that marketing is not the same as advertising; too often, they have a narrow understanding of it and are then easily molded into the perverse industry standards, ending up in the purgatory of meaningless campaigns while convincing themselves they’re doing something of real value.

But marketing is not about campaigns, and it sure as hell is not about “creating Facebook competitions”. Rather, marketing is a process of continuous improvement of the business. Yes, this includes campaigns because the business cycles in many industries follow seasonal patterns, and we need to communicate outwards. But marketing has so much more to give for strategy, if only marketers would stop wasting their time and instead focus on the essential.

Now, what I wrote here is only based on anecdotal evidence arising from personal observations. It would be interesting, and indeed of great importance, to find out if it’s correct that most marketers are wasting their time on petty campaigns instead of the big picture. This could be done for example by conducting a study that answers the questions:

  1. What do marketers do with their time?
  2. How does that contribute to the bottom line?
  3. Why? (That is, what is the real value created for a) the customer and b) the organization)
  4. How is the value being measured and defended inside the organization?

If nothing else, every marketer should ask themselves those questions.

Joni

Facebook Ads: remember data breakdowns

english

Here’s a small case study.

We observed irrational behavior from Facebook ads. We have two ad versions running; but the one with lower CTR gets a better relevance score and lower CPC.

This seems like an irrational outcome, because in my understanding, CTR as a measure of relevance should be largest impact factor to CPC and Relevance Score.

Figure 1  Aggregate data

So, we dug a little bit futher and did a breakdown of the data. It turns out, the ad version with lower aggregate CTR performs better on mobile. Apparently this adds emphasis to the algorithm’s calculation.

Figure 2  Breakdown data

Lesson learned: Always dig in deeper to understand aggregate numbers. (If you’re interested in learning more about aggregate data problems, do a lookup on “Simpson’s paradox”.)

Joni

On online debates: fundamental differences

english

Back in the day, they knew how to debate.

Introduction. Here’s a thought, or argument: Most online disputes can be traced back to differences of premises. I’m observing this time and time again: two people disagree, but fail to see why. Each party believes they are right, and so they keep on debating; it’s like a never-ending cycle. I propose here that identifying the fundamental difference in their premises could end any debate sooner than later, and therefore save valuable time and energy.

Why does it matter? Due to commonness of this phenomenon, its solution is actually a societal priority — we need to teach people how to debate meaningfully so that they can efficiently reach a mutual agreement either by one of the parties adopting the other one’s argument (the “Gandhi principle”) or quickly identifying the fundamental disagreement in premises, so that the debate does not go on for an unnecessarily long period. In practice, the former seems to be rare — it is more common that people stick to their original point of view rather than “caving in”, as it is falsely perceived. While there may be several reasons for that, including stubborness, one authentic source of disagreement is the fundamental difference in premises, and its recognition is immune to loss of face, stubborness, or other socio-psychological conditions that prevent reconciliation (because it does not require admittance of defeat).

What does that mean? Simply put, people have different premises, emerging from different worldviews and experiences. Given this assumption, every skilled debater should recognize the existence of fundamental difference when in disagreement – they should consider, “okay, where is the other guy coming from?”, i.e. what are his premises? And through that process, present the fundamental difference and thus close the debate.

My point is simple: When tracing the argument back to the premises, for each conflict we can reveal a fundamental disagreement at the premise level.

The good news is that it gives us a reconciliation (and food for though to each, possibly leading into the Gandhi outcome of adopting opposing view when it is judged more credible). When we know there is a fundamental disagreement, we can work together to find it, and consider the finding of it as the end point of the deabte. Debating therefore becomes a task of not proving yourself right, but a task of discovering the root cause for disagreement. I believe this is more effective method for ending debates than the current methods resulting in a lot of unnecessary wasted time and effort.

The bad news is that oftentimes, the premises are either 1) very difficult to change because they are so fundamentally part of one’s beliefs that the individual refuses to alter them, or 2) we don’t know how we should change them because there might not be “better” premises at all, just different ones. Now, of course this argument in itself is based on a premise, that of relativity. But alternatively we could say that some premises are better than others, e.g. given a desirable outcome – but that would be a debate of value subjectivity vs. universality, and as such leads just into a circular debate (which we precisely do not want) because both fundamental premises co-exist.

In many practical political issues the same applies – nobody, not even the so-called experts, can certainly argue for the best scenario or predict the outcomes with a high degree of confidence. This leads to the problem of “many truths” which can be crippling for decision-making and perception of togetherness in a society. But in a situation like that, it is ever more critical to identify the fundamental differences in premises; that kind of transparency enables dispassionate evaluation of their merits and weaknesses and at the same time those of the other party’s thinking process. In a word, it is important for understanding your own thinking (following the old Socratean thought of ‘knowing thyself’) and for understanding the thinking of others.

The hazard of identifying fundamental premise differences is, of course, that it leads into “null result” (nobody wins). Simply put, we admit that there is a difference and perhaps logically draw the conclusion that neither is right, or that each pertains the belief of being right (but understand the logic of the other party). In an otherwise non-reconcialiable scenario, this would seem like a decent compromise, but it is also prohibitive if and when participants perceive the debate as competition. Instead, it should be perceived as co-creation: working together in a systematic way to exhaust each other’s arguments and thus derive the fundamental difference in premises.

Conclusion. In this post-modern era where 1) values and worldviews are more fragmented than ever, and 2) online discussions are commonplace thanks to social media, the number of argumentation conflicts is inherently very high. In fact, it is more likely to see conflict than agreement due to all this diversity. People naturally have different premises, emerging from idiosyncratic worldviews and experiences, and therefore the emergence of conflicting arguments can be seen as the new norm in a high-frequency communication environments such as social networks. People alleviate this effect by grouping with likeminded individuals which may lead into assuming more extreme positions than they would otherwise assume.

Education of argumentation theory, logic (philosophy and practice), and empathy is crucial to start solving this condition of disagreement which I think is of permanent nature. Earlier I used the term “skilled debater”. Indeed, debating is a skill. It’s a crucial skill of every citizen. Societies do wrong by giving people voice but not teaching them how to use it. Debating skills are not natural traits people are born with – they are learned skills. While some people are self-learned, it cannot be rationally assumed that the majority of people would learn these skills by themselves. Rather, they need to be educated, in schools at all levels. For example, most university programs are not teaching debating skills in the sense I’m describing here – yet they proclaim to instill critical thinking to their students. The level and the effort is inadequate – the schooling system needs to step up, and make the issue a priority. Otherwise we face another decade or more of ignorance taking over online discussions.