Back in last October, I gave a talk at MapCamp, and part of the talk was basically me using a set of Wardley Maps to share a set of reckons, about why Facebook would have spent the the equivalent to the GDP of Iceland (i.e. NINETEEN BILLION DOLLARS) buying WhatsApp back in 2014. I did this to show how you can use Wardley Mapping to help understand previous strategic moves, so you know what to look for in future. Without further ado, here’s the expanded write up.
Before I start, I need to be explicit – I have never worked at, or for any of these companies – this is educated guesswork on my part, and basically punditry with maps. I’ll share the deck, at the bottom and I’d welcome alternative explanations here, depicted with maps, or some of my assumptions challenged here using them.
Anyway. Let’s go.
Starting in with a value chain
Let’s start with a value chain first to help understand the need I’m using Facebook to meet.
At the top is the anchor here, me. I have a need, to stay in touch with my friends and family. There are various ways I can do that, because I’m (juuust about) a millenial, I don’t really like phone calls. and I prefer using messsaging – i.e. email, SMS messages, and so on whatever. For simplicity, I’m restricting this map to show I might use Facebook for this, and I represent this activity as messaging. Of course, I need at at least one or more friends or family members to message, otherwise I’m not going to have a very fulfilling experience, and I’ll need to have enough spare attention to spend writing to them.
Now, I could obsessively hit F5 on my keyboard staring at an inbox somewhere (lots of social media apps email you when a message has been sent to you), but one of the key things that Facebook offered very early on was notifications in my browser when new messages came through, as long as I was on the Facebook website. So, let’s list those too in our value chain.
Now to be in touch with my friends over Facebook, I need them to be on the website, so lets assume I’ve invited them, and they’re represented using the social graph. While I’m paying attention to Facebook to hear from my friends, they’ll also siphon away some of that attention to direct it on ad inventory bought by advertisers – this is the deal most of us typically think we’re striking with them when we use the service.
Finally, because me and millions of other people are all using this one website at the same time, it requires a non-trivial amount of compute and persistence (i.e. storage) to store the messages, what ads I’ve looked at or clicked on, and all the other things that might be useful for Facebook to track that I might not be thinking about so much. The most common way to provide compute and persistence is by having a datacentre full of servers, so lets add that too.
Adding the mapping axis
We’re using mapping to explain this, so lets add the evolution axis now, as if it was the mid 2000’s.
Messaging isn’t really new by now, but in the mid-late 2000’s Gmail was still kind of cool and interesting product, and people still downloaded and used Thunderbird. Let’s put it it in the product/rental section for now.
My friends and family are definitely not an interchangeable commodity as is my friend list/social graph that’s generated from them, so let’s have them on the far left. Also, the Facebook website is not the first website ever, but it’s definitely a custom built one, as there’s loads of tech that’s being built that will eventually resurface as open sourced projects a few years later.
Ad inventory isn’t a new concept, and Facebook’s business largely relies on making it as easy to create ads, in a self serve model. Let’s put those to the right. When ads are this easy to make, becoming an advertiser comes more commonplace, so again let’s put that only a little bit to the left of the ads (you would expect there to be more ads than advertisers…).
Finally, towards the bottom of the map, Facebook manages to compete with others by buying loads of commodity servers and storage, but it has some custom built software to orchestrate and run their own infrastructure.
I’ve represented that by having them in the middle – I’m not really trying to account for their open compute project here, as it’s not the focus of this map,
Visualising how Facebook makes money
Let’s try visualising the flow of value here to help explain the the next few views of the map.
I’m meeting my need of being in contact with friends and family by directing my attention on the facebook website, which relies on notifications keep me engaged, and the closely guarded social graph (although, as we’ve found out recently with Cambridge Analytica, maybe not that well guarded after all). I’m representing this flow of value with the blue stroke along these lines.
As I use the site, part of my attention is being directed towards ad inventory which is how Facebook generates revenue. I’ve used green strokes to represent my attention being turned into cash money.
So far, even if you’re not that familiar with Wardley Mapping, I’m hoping this map shows something recognisable as a social media website that makes money by monetising eyeballs through advertising.
Adding movement and evolution
Now if you wanted to compete with Facebook here at this business, in the mid/late 2000’s you’d probably need to be able replicate all the bits I’ve shaded in grey here, and match Facebook’s own investment in:
- custom technology, for the user facing website site, and ad-serving infrastructure
- getting people to know the site exisrts, and then get them on the site
- getting people to share their social graph, to get others on the site
- building the actual infrastructure and including software to make it possible to store all the photos, messages and activity
As far as moats go, this is a pretty big one, especially as network effects here tend to favour the biggest player in a given market – given the choice between joining two networks, Facebook and say, Network B, people will often choose to join Facebook, because it already owns the social graph – you join Facebook because your social graph is already on Facebook.
How mobile changes this
In the late 2000s’ though, with the rise of mobile we saw some changes. I’m going to use Benedict Evan’s quote here, as he summarises the changes they bring about pretty well:
Smartphone apps can access your address book, bypassing the need to rebuild your social graph on a new service.
They can access your photo library, where uploading photos to different websites is a pain.
They can use push notifications instead of relying on emails and on people bothering to check multiple websites
Crucially, they all get an icon on the home screen.
Got that? Alright, let’s have a go at mapping this
Lets’ try to keep this simple, and focus on notifications.
Rather than needing all of Facebook’s infrastructure, and social graph to get notifications now about messages from your friends, you can think of notifications from friends as a feature available to almost any smartphone app that can access your address book.
So, it pushes notifications on our map waaay out to right instead.
Actually, we haven’t represented the address book as a more common alternative to Facebook’s graph, so lets put that on the map too:
Here, I’ve represented mobile apps, and smartphone messaging apps in particular as one of the alternative consumers of my attention.
For Facebook, this is bad news, as if there’s less attention being spent on the website, then there’s less to divert towards ad inventory, meaning less revenue.
I’ve represented the threat presented by mobile as the lines with the red stroke, and the withering flows of value toward Facebook, and advertisers as the much thinner green and blue lines.
This isn’t official, canonical Wardley mapping notation now, but for the purposes of this blog post, I hope it helps.
Responding to this threat
Back in 2010, I was in a consultancy, Headshift that was acquired by Dachis Group, a doomed VC-backed US consulting firm that had been hoovering up consultancies all around the world, in a bit to own ‘social business’.
After the acquisition, the plan was to get everyone feeling and acting like they were in one company, – so get everyone together in Las Vegas, for a multi day workshop. At the time, loads of Americans in this new company were going nuts over Beluga, a nifty new messaging app that they used to stay in touch with each other instead of Facebook Messages. Why I mean, use Facebook when this did a better job?
A few months later in 2011, this was the homepage for Beluga. Facebook had used its war chest to acquire the company – the product was integrated into Facebook, and became what we now know as Facebook Messenger.
Repeating this pattern with Instagram
You can see this happen again a couple of years later with Instagram, with photo sharing. Facebook acquired Instagram for around billion dollars in 2012, or around 1% of its market capitalisation at the time.
When you look at at Instagram by itself, spending such a huge amount for what at the time was as service with zero revenue, and less than 20 employees, it seems insane. When you look at is as heading off a fast growing threat to Facebook, it makes a little more sense.
Repeating this pattern again with WhatsApp
In 2014, WhatsApp had become popular messaging tool, and was growing fast – it had around 450 million users, and outside of China was probably the biggest rival to Facebook’s own messenger.
Viewed through this lens, you can follow the logic of buying a fast growing competitor to neutralise the threat to your business, especially if it’s something you’ve managed to do successfully twice before.
I think this is why they did it. It’s still a mind boggling amount of money to buy a small company for, and at the time, it represented about 10% of Facebook’s market cap at the time, so this wasn’t a decision taken lightly.
Did it pay off?
This chart below, from an article on Recode in February shows how many people were using these former rivals to Facebook’s dominance in late 2017.
Of course, these other companies have now grown this fast because of the huge wheelbarrows of cash brought in through the being bought by Facebook, but it’s not hard to imagine a scenario where had they not been bought, WhatsApp, Instagram, and Beluga would have grown to be threats to Facebook’s dominance.
So, are you saying Facebook is using Wardley Maps?
I’m using maps to help explain what I think happened.
This is venturing dangerously into full on armchair punditry territory now, but I think Onavo probably played more of a part helping Facebook identify threats early on.
Onavo is an Israeli firm that produced an app for iOS and Android that would tell you which apps were using all your data allowance, and as I understand it, reduce the bandwidth used to send this data to app servers.
So, you install the app, you save money on data, and and see which ones are rinsing your data plan.
In return, Onavo gets a really rich picture of which apps are being used the most on a phone, and where users as spending time, which they would use to create a product called Onavo Insights, showing which apps are gaining market share, and where the engagement is.
In 2013, Facebook bought Onavo, after being a customer of theirs for a few years.
Shortly afterwards, Onavo Insights, the market intelligence product, went off the market.
I have no idea if there’s an internal version of this in use at Facebook now, but if you did want to identify competitors on mobile to acquire before they got too big and dangerous to your business, it sure would be handy if you were the only one who had it, right?
Come at me with your maps, cuz
As I said, this is at best educated guesswork, but I hope it’s at least been interesting.
I’d really like someone to tell me they disagree, and show me how these maps should look.
So, I’ve shared the deck with all the maps below.