Improving conversions is easy. Improving long-term engagement is hard.

It comes down to finding a well-defined proxy that contains lots of hard data and allows you to do fast experiments.

We’ve been doing Conversion Rate Optimisation in SimpleSite for a long time and we have found a large number of wins in conversion from traffic to paying users over the last couple of years – between 5% and 20%. Needless to say, that makes for a tremendous difference in growth and a lowered Customer Acquisition Cost (CAC ) for new customers. Lately we started to focus more on increasing engagement. We believe that by increasing engagement – or Customer Success – in the first few minutes of the user experience, we can gain a next set of big wins. Better engagement should give us all of the below

  1. An increase in the number of conversions from trial user to premium, leading to more paying customers and a lower CAC.
  2. An increase in the lifetime of every converted customer, if they become better engaged and experience more value in the first half hour of their lifetime with SimpleSite. That should lead to a higher average LTV for the converted customers.
  3. Better engagement for users that continue as free users in our Freemium business model. That should lead to more virality AND a higher chance that they convert to premium users at a later stage in their lifetime with us – effectively raising the conversion rate and lowering CAC, but at a much later time than our current initial wave of conversions in the first days of the customer lifetime journey.

In all, a number of effects that improve growth and the LTV /CAC ratio, which is a key parameter that determines the profitability of the individual customer, sets the parameters for paid acquisition and therefore ultimately affects growth in a substantial way. So far, all seems well. Of course, one should increase engagement. So why not just do it? Why should it be hard?

Why is improving engagement so difficult?

Let me start somewhere else instead. I’m implicitly saying that Conversion Rate Optimisation (CRO) is easy, at least compared to optimising engagement. How so? Let me quickly say that when we are talking about improving any metric we mean improving in a quantitatively verifiable way. So we need to be able to do an A/B experiment showing us that we did something better than before. Now this is why I’m saying that CRO is (relatively) easy: You have lots of data. You have a completely well-defined success criterion. It happens fast. Or, a little more detailed

  • CRO happens at the beginning of your funnel where you have all the leads and quite a lot of conversion events. So there is volume, which you need for statistics to make sure your findings are significant.
  • The obvious success criterion for CRO is “do they convert?” – no hesitation in what to measure.
  • And it happens in days or weeks. You don’t have to wait years to see whether a potential customer does the desired thing.

How does that relate to optimization of engagement? What is real engagement, truly, in your service? That is for you to decide, but it probably relates somehow to long-term retention or high lifetime value.

How long can you wait for results?

And therein lies the problem. Ideally, you would want to do an A/B experiment on some onboarding idea and then half a year later test whether you got more retained users out of it. Half a year is a long time to wait. And probably, rather few of your users are really that active after half a year, especially if that is what you are aiming to improve. So you wait a long time to measure on rather few users, which means that you have to wait even longer to get enough data for significance. Not exactly a recipe for testing many ideas in a fast loop of idea-measurement-new idea-mesurement (repeat).

Why improving conversions is easy and improving long-term engagement is hard. #growthhacking #engagement

So what do you do? Find a proxy

And here is where the notion of Activation from Dave McClures AARRR framework comes in very handy. You are really looking for a short-term proxy to use instead of long-term engagement. A short-term proxy that

  • Only takes a few days to completely record for a given user
  • Has lots of data in it
  • Relates to long-term engagement in a meaningful way, so if the proxy increases, long-term engagement increases

That sounds a lot like Activation. Our experience is that if you can’t measure something fast and if you don’t have lots of data, you can’t empirically improve it. But with fast data and lots of volume, you can really move the needle. And if you look at the finest in the Growth Class, they all have that parameter. Facebook has their seven friends, LinkedIn has X connections in Y days and Zynga looks at users retained on day 1. And they all emphasise that particular insight, because it allowed them to systematically improve things.

How to find the proxy

Now that is the art. What the actual proxy should be is not a given. Remember, it has to be fast, contain a lot of data and predict long-term engagement in your service. The proxies that you read about out there are typically of some threshold nature. Users must do some action X times in the first few days to count as a success. What that action is and how high an X is needed is particular to each specific service and its users. What proxy have you found in your business? Share your results here at Nordic Growth Hackers. I’ve shared SimpleSite’s “proxy” at the post Does SimpleSite have a “7 Friends in 10 days” like Facebook?

Morten Elk

CEO at SimpleSite
CategoryExperiments