David Low, November 2015
Making Products, Making Money
• Founded 2003
• 800 employees
• 11 international offices
• 350+ direct airline partnerships
• >400 Online Travel Agencies
• Flights into Travel: Hotels, Car
Hire, Research & Insight
• 80% of revenue outside the
Over 50 million
Over 35 million app
Over 30 languages
Over 300 partners
use our Affiliate and
35m 30 300
Some headline figures
We want to ‘power the travel internet’.
Affiliate schemes have been around since the Internet
• “You grow my business, I’ll give you a small share”
• Brokers like TradeDoubler and Commission Junction dominate
the market – keeping it all honest
• Over the year it’s been hard to use a blog without an array of
buttons littering the screen.
• But also over the years these models have got larger and far
more sophisticated as we’ll see later.
Nothing changes (part 1)
The original problem
Sales Marketing Product
Two sided markets
• Wikipedia defines this as:
“Two distinct user groups that provide each other with network
• This is the very definition of an affiliate model. Two sides working for some level of
mutual benefit – with a platform mediating between them.
The pure defitinion of an affiliate model
• Probably the best example of an affiliate / two
• Apple creates the conditions to build and sell
apps; developers create them.
• In return for making a better use case to have
iOS devices, Apple pays roughly 60% of sales
fees back to the app publisher.
• $42 billion generated in around 7 years - $25
billion of which went to publishers.
iTunes App Store
• Everyone knows what AdWords and AdSense are
• Effectively an affiliate model – advertisers and
publishers meeting in the middle, associated by
Google who pay out a share.
• Up to 75% of revenue is given out to publishers,
in return for enlarging Google’s audience and
opening up more data
• $14 billion a year paid out to ‘network partners’
AdSense and AdWords
Affiliate payment models
• The two examples were chosen deliberately as they show the two most common types of
• CPC (cost-per-click) is one method – you put something in your product, every click is
tracked and pays you something back. Google Ads are clearly on this model, although the
‘click’ price is figured out by the two-sided market.
• CPA (cost-per-acquisition) is another – you put something in your product, clicks are tracked
but only paid if an action is completed (usually a purchase). Apple is clearly this model.
• CPM (cost per ‘mille’ / thousand) is a model most used in display advertising, but not
These two examples are illustrative
Sitting between two-sided markets
Publishers Skyscanner Providers
Skyscanner and Providers
Nothing changes (part 2)
• Going back to Geocities and my original blog 20
years ago, the problem was attracting traffic.
• That’s not such a big problem now, traffic can be
acquired for money – but you have to ensure the
value generated, exceeds the cost of acquisition.
• Clearly retaining that user for longer, with a
‘lifetime value’, has a greater chance of
exceeding the cost.
Acquisition and retention
Doing your marketing for you
• By giving other brands and products access to your data or services, you open up far more
users than you might get on your own.
• There’s a good chance you will inherit some of the loyalty shown by those users to their
• In addition those brands might extend the use cases for your product or data, beyond
something you had planned or expected.
• We say we want to “power the travel Internet” and this is exactly where it comes in.
• Some good examples of Skyscanner affiliating with other brands…
This is where affiliate marketing comes in
FlyBe: Additional revenue
New Ideas: Hitlist
The APIs are
lightweight and easy
to integrate with, and
it’s our primary
Gillian Morris, CEO Hitlist
Create a product suite
The Skyscanner for Business Suite
The more you open up, the more people can innovate
Think about the whole market
Sales Marketing Product
It’s all about good products on both sides
• All the examples above are about good products
• Good products on both sides of the market add up to a lot
• User-facing product owners need to think differently to
acquire and retain users
• Affiliate providers need think about the whole market and
innovate to help their clients
Thank you – questions welcome
Skyscanner for Business Website
Skyscanner for Business, Quartermile One, Lauriston Place, Edinburgh EH3 9EN
Affiliate models are as old as the economy. You help me do something, I’ll give you a share. Grow my business and I’ll pay you a dividend. Geoff promised us 5% of the ticket money from tonight - before I realised it was free.
The commercial Internet was built on affiliate schemes. People of my vintage thought they would make their millions by robbing some web space on Geocities, getting noticed on Alta Vista and sticking loads of “buy me now” buttons on a page.
The problem was the ‘getting noticed’ bit. Even in a web with a few thousand pages, acquiring users was a bigger problem than monetisation - something I’ll come back to later on.
With respect to my friend from Amazon, which has something like 900,000 people in its Associates programme - the biggest two-sided networks on the Internet are probably Apple’s App Store and Google’s advertising programmes.
Apple started out with an ecosystem. It provided a means of putting apps in a store and means of users quickly finding them. The store has largely got out of hand in recent times and favoured those with high marketing spend, but the theory is largely the same.
This is all effectively a classic economic model called a two-sided market, or a two-sided network. Wikipedia defines the two-sided market as such:
“Two distinct user groups that provide each other with network benefits”
It’s a relatively simple concept but in practice it’s a step beyond the Geocities example above.
The important part of the equation is the ‘platform’ in the middle. Without that platform to facilitate the market, like any stock exchange, there’s no market. Like my old web pages, there was no interaction between me and the affiliate partners to make things work.
Developers create apps, they charge for the apps or the content inside them. Their time is rewarded by users who pay, and in turn that creates better apps. Apple sits in the middle taking around 40% of the money - which pays for the ecosystem and obviously profit.
Apple claims it has paid out over $25 billion to developers over the years, which means the market overall has generated $42 billion. But ultimately it created something many people can’t live without - on both sides of the market. Apple is doing pretty much anything it can to ensure that native apps are the dominant use case on its devices.
Google’s market isn’t seen as an affiliate network but it’s a very capable two-sided market. Google clearly provides the platform for buyers and sellers of advertising.
Publishers provide space to put Google’s ad slots in, Google finds people to buy them. It’s not the traditional affiliate model but in return for making Google’s network bigger, they give you a slice of the money. Approximately $14 billion dollars per year, at a share of almost 75%, is paid out to ‘distribution partners’.
I chose those two examples very carefully because in effect they demonstrate the two primary models of how affiliate marketing works.
The first is what’s called “cost per acquisition”. Apple sells your app, you get a large slice of the money. Amazon sells a book from your referral, they pay you a percentage.
Many other sites operate both models very successfully - look at the success of meta and comparison sites - MoneySupermarket, Compare the Meerkat, Go Compare and so on - and discount aggregators like Quidco and Top Cashback. The latter operate by getting some kind of commission for a sale, and giving the user a share.
Skyscanner falls into the ‘meta’ category - a comparison site which searches on your behalf. Clearly I’m not going to tell you Skyscanner’s precise business model - I might not have a job tomorrow if we give that away.
But with all I’ve said so far, you can see Skyscanner is in a relatively unique position to talk about the challenges attached to this kind of activity.
Where I work in our B2B team, we effectively live in between a pair of two sided markets. Let me explain this a little bit.
Inside Skyscanner we like to talk about the ‘flywheel’ of our core business. We make our site more attractive to our customers, we generate higher-quality traffic.
In return our providers make our life easier in providing data and ‘content’, because that’s ultimately what our site shows. Which in turn produces more, or retains more, users who want to buy travel. It works in a constant circle and both sides of the market speed things up.
The problem with this is, we aren’t selling anything directly. We do our job and pass people onto airlines, hotels, car rental companies. We hope the service we provided - and our ‘flywheel’ - will bring those users back next time, but sometimes they won’t.
So our problem is not only acquiring those users - it’s retaining them for next time, and for a long time after that.
Retention is a constant battle for every site - you pay for marketing to acquire users, and need them to generate more money than you pay.
This is where affiliate marketing comes in. By giving your content away to other brands, sites and partners, you acquire their users in some way. Retention is all down to the other site’s brand loyalty and thus out of your hands - and less of a problem.
So at the same time as operating our own site, we work very hard to ‘power the travel internet’, as I mentioned earlier. Effectively helping to create other travel products that cater for different contexts, situations, brands and audiences which have their own ways of acquiring and retaining users.
In return for replacing your acquisition cost, you pay some of the revenue you make back to the other site.
To illustrate that, I’ll throw in a few recent examples.
We recently announced we are powering a section of the MSN site. This is a classic affiliate product called a ‘white label’ - we’re effectively taking our own site and re-skinning it for another brand. Again we’re acquiring someone else’s traffic.
Clearly MSN is one of the bigger brands on the Internet, to this day - reaching a loyal audience of 100 million users. They wanted to keep their core products MSN branded so that’s a really good fit for us.
FlyBe wanted to be able to offer flights beyond its own network, because of relationships it has with other airlines.
We enabled this by building their “One Stop” product, which lets users see where they can go after an initial FlyBe flight. It extends our user base into loyal FlyBe customers whilst making FlyBe’s product better.
GoEuro have a slightly different use case. As an end-to-end (“Multi Modal”) travel search, they need the best coverage of flights to make their product better. We enhance their product and again, attract a different class of user than people who might only look for one form of transport on Skyscanner.
This one’s a little bit different - Hitlist is an app trying to disrupt the inspiration to travel. Their motto is “acquire experiences rather than things” - spend your money on memories.
So it’s very different from the comparison model, it’s more of a magazine or suggestion format, but of course it has Skyscanner inside.
This is my pet project just now, enabling search by voice. You ask something of the speaker - or anywhere else Alexa’s service is embedded - and you get your answer back. It’s a natural home to get travellers thinking early and we’re very excited about it.
I deliberately put these examples in this order because it shows the different kinds of thing we get up to - the last two is an example in reverse. Amazon make it so easy to work with their platform, we’re naturally drawn towards it.
Amazon don’t look for a share of anything we make on that platform but if we make it more attractive, more people will buy it. Another two-sided market we’re attacking from a differnent angle.
For us it’s all about a product suite. Offering lots of flexible tools, well supported and documented, for people to adopt and extend. None of the examples I’ve just been through, would exist without our B2B product line. It’s hard to overstate how important they are to us.
But more than that, it’s about attitude. We had an experiment a couple of months ago where people from one discipline had to present for 20 minutes on a different one. For my sins I had to present on Marketing - but at the end of the day, by helping people to build things with that product suite, it’s a modern form of marketing.
I could just as easily be in sales or product - it’s all about understanding all parts of this two-sided market to get somewhere.
I hope that all made some sense and to sum up, I’d underline these points: