WMF as an advertiser

Noto Emoji Pie 1f4c4.svg (English) This is an essay. It expresses the opinions and ideas of some Wikimedians but may not have wide support. This is not policy on Meta, but it may be a policy or guideline on other Wikimedia projects. Feel free to update this page as needed, or use the discussion page to propose major changes.

Wikipedia, has no advertisers. While that may be true for external advertisers, a case could be made that WMF is its own biggest advertising client. WMF can advertise on top of any page, in any country, on any project, as much or as little as it wants. The donation/fundraising campaign that is suited to most non-profits or even online properties, is slowly turning in to an advertising campaign renewed every year, to generate the maximum click through with little underlying concern for the message.

WMF fundraising department, has one primary source of revenue. The banners on top of every wikipedia page. The team has been working diligently for the last few years "A/B testing" everything under the sun to evolve their message and maximize as much money as they can to get out of every viewer. To do this, they have had little sight of the project and principles that Wikipedia was based on.

WMF as an advertiserEdit

If for a moment, we consider WMF as just an advertiser, and hold them up to practices that most reputed advertisers go by -

Campaign specific messagesEdit

  • "Urgent appeal from Jimmy" - This sort of target messaging is usually reserved for actual emergencies and disaster relief. On clicking through one would expect to read about a Flood in India, a disaster in South America, or some tragedy in Sudan. This emotionally manipulates and misleads the reader in to the urgency of the message that follows. Most people who donate might be horrified to learn that the actual money they donate would go to developing unwanted things like Flow and Moodbar.
  • "If everyone reading this donated X fundraiser would be over now." - Yes, if everyone in an entire country donated(or 20%), our target would be met instantly. If Coke or Pepsi had the same banner about drinking their products, they would hit their quarterly sales target instantly too - kind of like how the fundraising team looks at it - with much worse banner art and design.
  • Banner placement and design - It's almost as if the banners are made to look amateurish. Black outline, greys so they can contrast with bright screaming fonts in yellow and red. Given how much the fundraising department has at their disposal - between specialized staff, designers and outside contractors - it is curious to know why they tend to not have a professional looking banner, they were evolving towards it a couple of years ago with Jimmy appeal banners but then stopped all together. I believe that an amateur-ish look makes it seem more real that the community is asking for this donation, not a professional staff of dozens with outside consultants, designers and millions at their disposal.

A/B testing the worldEdit

A/B testing has become the bane of our times. As one of the most used memes pervading the wikimedia ether, it is reminiscent of a time of, "a data driven approach" overpowered common sense, decency and sanity. Yes, A/B testing establishes correlation between two different perspective. Yes, the results are comparable to a certain extent. But it is all relative. In the simplest form, A/B testing leaves out C,D,E,F and all the other alphabets or the english language or common sense in some cases.

There are far too many variables to simplify two results in to two related causes. The correlation there can arbitrary, tenuous, incidental or completely misleading. Some degree of common sense has to be exercised about choosing two narratives to test against each other or their results. The banner team has used A/B tests to justify most of their decisions, the more they relied upon it, the more they became blinded to everything else, and further they have gone from acceptable standards for fundraising.

Why does a small change yield this result over another. It is oversimplifying variables that we might not be able to perceive or quantify. At its core, it is establishing relation between two moments in time. It is a rhetorical question to ask why one moment is different from another, or one day from next. Philosophers, pontificators have been wondering for all of humanity though they never claimed what they knew as "data driven".


This inevitably leads to where do we go from here. The banner department trying to beat target every year and expanding 40-50% or more over each year. How can they continue increasing their target? Here are some thoughts that might have already been considered -

  • A splash screen over WP main page just to get some data. Or better yet, a pay wall that can only be bypassed by donating or performing some ritual to appease the banner lords.
  • How about hiring the true banner experts of the world - Spammers. Be it Porn, Viagra and Gambling. They have been at it for decades through the same banners. I'm sure their appeal, and click-through rates would leave the fundraising team behind.

Further A/B experiments -

  • Team wearing white instead of black has any effect on targets.
  • Facing North vs east.
  • Drinking Pepsi or coke.
  • Praising Cthulhu

It should occur to someone (eventually), the correlation there is spurious. Kind of like the relation between Films Nicolas cage appears in and the number of people who drowned by falling in to a pool. Minor incremental changes to banners yield minor discrepancies in revenue, while everyone would like to believe there is an established correlation between the two, there are far, far too many variables to account for. Anyone intelligent enough would be the first to admit that "A/B" testing is leaving out the C,D,E and just about every other alphabet.

It is time to step back and look at the larger picture, maximizing page views isn't a science. If it was, porn, viagra and gambling banners, the true spammers would be the biggest expert at page view maximization and click through rates- the true data driven scientists of their field. I'm sure they based their conclusions on a/b testing everything.

Judge ourselvesEdit

If we were to evaluate ourselves as advertisers, how would we be rated? - Ethical? honest? or would we be in front of some judge or at the target of consumer group, for false and misleading advertisement? Or is it acceptable to just stay at the boundary of being almost unethical and illegal in our advertisement?

Shouldn't we have higher standards? As one of the leading source of educating the public, and disseminating information, shouldn't we value the same things Wikipedia values - Truth, Verifiability, Neutrality. No amount of fundraising target can fix the core problems. They are however, making us more and more dishonest and taking us further from our true identity.

The next question to ask is, Why are we doing this? So, we can have another failed unwated extension or 20 new staff members. Clearly all the attempts at fixing the real problems have either failed, outright or only generated limited results. From increasing diversity in terms of "Global South" or female participation, to heading of some exodus of editors or readers. It is arguable which of those problems needed fixing, or if they are fixable - they might just be the symptom or the price we pay for being......us.