Tuesdata: The most interesting data from Facebook’s infamous Cunningham Memo
Welcome to Tuesdata, where the hogging of the headlines around the world by Meta continues, and Unmade is happy to help that trend continue.
This week we will pick apart an internal memo from Meta that late last month was made available to the public thanks to the U.S. congressional antitrust subcommittee. It didn’t receive a significant amount of coverage, particularly in Australia, which is a shame as it contains some very interesting information.
The full content of this post is for Unmade’s paying members only. Others will hit the paywall a little lower down.
The short version of this story is that the US Congress is considering antitrust legislation. The legislation would make it more difficult for companies to stifle competition either through acquisitions or their behaviour.
Part of that push is called the American Innovation and Choice Online Act. The proposed bill was introduced into the House of Representatives on June 11, 2021, and has advanced to the stage where a floor vote is now expected in September this year.
The bill aims to prevent big tech companies from self-preferencing their products and services at the expense of the competition, particularly the smaller players.
Furthermore, in late 2020 the US Federal Trade Commission sued Meta in an attempt to try to get it to divest Instagram and Whatsapp.
There is a lot going on – this is a rabbit hole that would be easy to dig too deep into, so let’s stop there and instead focus on the aforementioned document that is part of all of this – what’s come to be known as the Cunningham Memo.
Written by Thomas Cunningham in 2018, then a data scientist at Facebook, it was created for Mark Zuckerberg to illustrate in depth the environment the tech giant was playing in. While Meta had been suggesting in public that it had myriad rivals, the memo suggests that it was more focused on itself rather than the competitors.
Not only that, it was keen to ensure that its brands such as Facebook, Instagram and Whatsapp didn’t have negative impacts on each other.
Below we will take a look at some of the most interesting aspects of the Cunningham Memo and pull out some of the key data points.
For those who wish to read it in full, you can download the Cunningham Memo here. It’s a short but worthwhile read, although some of the graphs and charts are particularly hard to see. Some of it is redacted, but only a small amount.
While it was written in the US fall (autumn) of 2018, there are some particularly interesting data points which are worth taking note of.
Here they are below as direct quotes from the memo.
5: The average amount of social apps used
The average mobile phone users use 5 different social apps in a month: they use an average of 3 mainly-messaging apps (such as WhatsApp, Messenger, iMessage, SMS, etc.), and 2 mainly-broadcast apps (such as Facebook, Instagram, SnapChat, etc.), out of 30 total apps per month. Every day they use 1.5 messaging apps and 1 social app, out of 10 total apps per day.
2: The number of sharing apps that can co-exist
While many countries can support two apps with high participation (10% or more of the population share on that app on a given day), it is rare to have three different apps which cross that threshold. A realistic family strategy should take this into account.
(Teenagers form a somewhat separate group.)
>50%: The percentage of the market messaging apps require to succeed
Most messaging apps which do not have 50%+ market share are shrinking. The apps which survive without majority-usage are typically attached to some other product/service. E.g., SnapChat (attached to Stories), Messenger (attached to Blue), Instagram Direct (attached to Instagram), iMessage (compatible with SMS).
Where Messenger is the majority-usage app, it does not have as high reach or engagement as WhatsApp in a similar position (e.g. Philippines, Iraq, Algeria). The extra time-spent on WhatsApp can be partly attributed to WhatsApp Status, but we also see that WhatsApp has a higher rate of messages sent.
Where WhatsApp is not the majority-usage app, it generally has very little presence. Most of the exceptions are SMS markets in which WhatsApp is growing rapidly and may become the leading app soon (e.g. UK).
>90% or <10%: The reach of typical messaging apps in a country
Typically reach is either above 90% or below 10%. This implies that these apps do not have stable equilibria with intermediate levels of reach, and so there exist tipping points. To be precise: if you were able to temporarily lower the reach of one of these apps from 90% to 10%, then it’s likely that, because you crossed the tipping point, that app’s reach would remain permanently low and another competitor app would take its place.
200: The number of friends the median monthly Facebook user has
The median monthly-active user has around 220 friends on Facebook, 100 contacts on WhatsApp, and 110 followers on Instagram (65 reciprocal followers). The small number of friends/followers on Instagram is partly attributable to (a) the lower reach of Instagram; and (b) the lower average time-on-platform, thus extrapolating current trends we expect Instagram’s graph to grow.
33.3%: Amount of original content from Facebook created by friends
Around 1/3 of Facebook feed is friend original content. But on Instagram it’s much lower: just 6% is from normal reciprocal follows, 16% normal non-reciprocal, 55% of feed is top-accounts; 25% is interests.
Flat: Engagement on Facebook at the time of this report
Engagement is flat in 2018 after falling in 2017. Original sharing is roughly flat. The increase in Stories has roughly offset the decline in Feed sharing.
Three things to think about
There are a number of interesting things to think about in the information that has now seen the light of day from the Cunningham Memo. They were most eloquently summed up by The Atlantic CEO Nicholas Thompson. He would know a thing or two about technology, having previously served as editor-in-chief at Wired.
Summed up, it’s that:
There is generally one dominant messaging app in each country. Pushing Whatsapp past 50% usage in a country is likely to mean it gets towards 90% very quickly. Fail to reach 50% and the chances are it will trend towards 1% very quickly. Therefore it is relatively easy for Meta to decide in which countries it should focus on pushing Whatsapp.
Teens should be thought of as their own group. Apps can succeed amongst teens and nowhere else, or everywhere aside from teens. They don’t follow the trends.
TikTok is likely to be a significantly increasing concern for Meta. The Cunningham Memo suggests that the increase in popularity of a social app doesn’t necessarily mirror the decline in popularity of the one it is about to usurp. Instead, the decline will occur after the competitor rises and is first seen through small declines in visitation and engagement. This article in the AFR shows just how much TikTok has risen in Australia and how much it has invested so far.
One step forward, two steps back
The Unmade Index ended last week on a slight positive, up 1.77% to 712.7, but it’s started this week with two consecutive falls to land back under 700, at 693.8.
The falls were led by a 5.32% drop from HT&E and a 4.33% decline from Ooh Media. HT&E’s radio counterpart, SCA, didn’t fare much better, recording a 3.78% fall.
Revealing all
Thanks for reading through to the end of Tuesdata. Hopefully the rest of Tuesday will be a successful one for you.
At the very least, there will be plenty to keep you entertained with the significant amount of tent pole programming that has begun to air on TV.
With The Block’s first reveals now done I’m likely to invest most of my TV watching time this week on seeing how the show progresses as it finally has a crack at a regional area.
As ever, if you have any feedback you can reach out on letters@unmade.media.
Stay safe,
Damian
damian@unmade.media
