As an EIR, I spend a lot of time talking to startups about their problems and how to solve them. I sat down recently with one of the Oxygen Accelerator teams to discuss the the biggest problem by far for any startup – customer acquisition.
Here’s an example flow of users for a generic mobile application:
We get an email address from a mailing list, send an email with a link, that link goes to a landing page, the user clicks a download link, doesn’t uninstall straight away, and does something that brings revenue. That’s an ideal scenario. The reality is that at each step, you get some percentage of people who convert to the next step, and some percentage that drop off. For example, if we email an app download link out, and 95% ignore the email while 5% get to the next stage and download, then this stage has a 5% conversion rate. Putting all the stages together forms a funnel. Our objective is to drive as much revenue add possible, and we do that in two ways:
- Improve the conversion rate at every stage
- Put more items into the funnel
Improving conversion rates
Your most dangerous weapon as a non-coder is a spreadsheet, so let’s unsheathe and wave it around. Here’s some sample data for the flow above:
Nothing radical, but it’s very important to keep and maintain your understanding of this funnel – edit it up as things change over time. A common question – how do you choose which events to put in the funnel?
- You have to be able to actually measure them. For example, you can’t measure views on your app’s page on iTunes, even though you might dearly love to.
- You should measure events where you have the ability to improve the conversion rates between them. Sometimes the conversion is out of your control.
- Pick the right level of detail for you. If you can handle email send -> email open -> images downloaded -> link clicked -> landing page opened, go for it – or you can keep it simple.
More users into the funnel
Ideally, users would just appear at the left side of the funnel, waiting to be sucked into your engine of revenue. This does happen – it’s called organic traffic. You put up a website – people search, find you, and pop on. You put up a mobile app – people search on the app store, find you, and ping! Users. There’s a bunch of things you can do to help (e.g. SEO) but we’ll leave that for another post.
Your other option is to pay to get users in.
There are hundreds of companies out there waiting to sell you ways to do this – your job is to pick the one that gives you the best return. How do you do that?
- First, calculate your customer LifeTime Value (LTV). The simple way to think about it is – what will an average paying customer actually pay me? If you are selling an app for $0.99, it’s $0.99. If you have subscription model, say $5 per month and you estimate your customers will stick around for one year, it’s $5 x 12 = $60. (For you pedants out there, I am ignoring the time value of money.)
- Second – and this is the fun bit – work backwards through your funnel to understand the value of a customer at each stage.
So let’s say your LTV is $60. It looks like you should be paying a maximum of $ for a download. But wait! There’s a rule of thumb here, which is that your marketing costs should be less than 50% of this figure. Therefore, you should aim for these as your marketing costs – 60 cents for a download, 1.5 cents for an email, and so on. Remember these are maximum costs – you should aim for the lowest cost possible that converts well.
Speaking of which …
This model is not perfect
One pretty model that solves the biggest problem your startup has? Not so fast! There are some problems to bear in mind which relegate this from “manna from heaven” to “useful tool” status.
Firstly, there is actually more than one funnel, because not all users are the same. Practically, this means that your conversion rates will differ for different types of campaigns. I’ll illustrate.
Let’s say you run a CPI campaign, where you are paying some amount (let’s say $1) for an actual install of your app. Think about the flow for those users – perhaps they’re in the middle of playing Dungeon Keeper, and the app asked them to install an app so they can get a free powerup in the game. Fact is, they’re not interested in the app, they are interested in the powerup so they can slay all sorts of grisly monsters. So the chance of them converting from a download to revenue is pretty slim.
Now let’s say you run another campaign where you go to a club for people who are all about the service your app provides, and install the app for them. (I know it sounds theoretical, but amazingly it’s not.) The chance of them converting from download to revenue is very high.
You can handle this by thinking hard about how similar your users are to each other for any particular campaign. If they are wildly different, you need to construct a new funnel for them and calculate using your new figures.
Second piece of bad news – in reality you’ll discover that the funnel isn’t exactly linear. For example, you can go from app download to revenue straight away, depending on how you are measuring retention. There’s no great solution to this – just try to keep your funnel as close to your understanding of the flow of real users as possible.
Finally, the funnel doesn’t take into account how things change over time. Many businesses are started by paying way over the odds for content or conversions (Yelp being a great example), with the understanding that acquisition costs decrease as the value of the service increases with scale.
Keeping an eye on your funnel is a critical activity for your business. Keep in mind the rough edges and you’ll find it essential for visibility of your business.
(Thanks to Jade at Sorted for a great discussion!)