Customer Love and Snakes & Ladders

One of the foundations of our business is what we call ‘customer love’. It all stems from the early days when we realised it’s much more fun and fulfilling to build software people want and love rather than trying to sell it! It’s a principle that’s built into everything we do from how we build the website, communicate with customers and structure the business.

But how do you measure it? A year ago Guy found a fascinating article by HubSpot, an American software company that tries to identify customers that are about to leave them through a measure called the Customer Happiness Index or CHI. At a certain threshold, customers are likely to stop using their software so they use it as a trigger to intervene and ‘save’ the customer.

At the same time, we had been looking at lead scoring systems like Marketo to help our marketing efforts and realised they are really the same thing. Lead scoring applies a similar principle – once a lead has a high enough score they are deemed to be ‘hot’ etc.

And there it was – our little eureka moment! If we could combine both of these into something that worked throughout the entire customer experience we could use it to formalise and embed the principle of customer love throughout the business.

Combining both ideas means we can start to get an impression of customer happiness right from day one. The next question though is much harder – how do you decide whether someone is happy and what do you do with that information?

Answer – play snakes and ladders!

Here’s the idea – we have a number of ‘games’. Each game has squares. A customer can only be on one square at a time and moves between them via the green ladders or red snakes, a move which is triggered by some event happening. A ladder gives you extra CHI points whilst a snake takes them away.

An example of a very simple game board is shown below:

Customers start playing this game when they register. For example, they move from the start square to the installing square and get 20 CHI points. 20 points because that’s a good thing and they must be reasonably happy to have registered. Next, if they create a license then they get more points (climb another ladder) and move on in the game but if they don’t do anything, they follow the red arc (the snake) and go down five points. You can see that the game is played in response to events happening (registering, creating a license, making a payment etc).

Now, we’re unusual because we build everything ourselves. From our website and marketing emails right through to the web service platform, the entire business is built on a massive chunk of in-house middleware. This is where we were incredibly lucky because with everything passing through this codebase we could build a large-scale eventing system that triggers an event when something interesting happens inside it. Our gaming system just listens to these events and moves the customers around these virtual games.

The games serve two purposes – we have a really simple way of seeing where they are in the game and the paths they have taken. A happy customer should be able to rattle through the website quickly, without any errors happening; then move on to make a payment and over time see a general increase in usage. The aim of the game is a steady increase in CHI with minimal drops. Where the CHI starts dropping off we need to understand what terrible things have happened. Where the CHI is bouncing up and down, and the customer is stuck on the same square, something’s gone wrong and they are probably struggling.


When we started playing the games we did so passively. In other words we used the games to follow behaviour and build information which we could analyse and understand. The next step was to become an active participant in the games. In other words – what can we do to intervene and improve the results for our customers?

These insights are invaluable. Where we see a group of customer ‘stuck’ we know there’s something there which isn’t working properly. We can then target our design and development resource into fixing the problem and easing that flow. Where we see the CHI starting to drop we can jump on it early to help out.

We’re now looking at being smarter too: we’re starting to analyse patterns in the games that indicate future problems. We’re trying to see if there are other indicators, not just in the games which we can use to pre-empt common issues and head those off early.

What’s great about this approach is that it’s very easy to visualise. We can look at a customer’s CHI chart and get an instant sense of where they are. We can even drill down to see all the events which contribute to a segment – an invaluable way of seeing what’s going on.

From a management point of view though, we can look at the current global status of the various games – how many pieces we have on each square! That’s a fantastic way of seeing what’s coming up and where we need to focus things. Plus it’s far more interesting than a classic sales funnel (yawn!).

Longer term we’re even thinking if we can take the principle further. Some of the interventions can be automated, for example an email to say ‘looks like you’re struggling with something, have you tried this’. But others obviously need the human touch. Those human interventions are the most expensive and we should therefore aim to get the biggest, positive impact on the CHI from it.

One option we’re considering is whether we can calculate part of our bonuses on the net CHI uplift we bring. For developers and designers we can try to code out the problems; for support and sales they can turn things around. And ultimately its single goal is to focus us all on creating a service and products people love.