Designing metrics for a start-up
Simon Gray, a data analyst at OpenAgent, explains what designing good metrics for a start-up is all about.
In this fast moving world of technology-enabled disruption, start-ups are working on so many new ideas there’s no telling what kinds of information they are collecting, let alone what they need. But one thing’s for sure, if they want a competitive edge, they need to be able to collect good data and to read it well.
Good data collection and analysis is vital to anyone who wants to know exactly where they stand. For a start-up up, it is even more important because they are trying to build something that has not been done before. This means they can’t just resort to common assumptions about the impact of changing something in a product or service.
At OpenAgent, we’ve made a couple of key changes to our sales process recently, with a focus on improving conversion rates that directly affect the bottom line. The insights we built these improvements off would never have been visible without good metric design.
So where do I come in?
When people who aren’t tech savvy ask me what I actually do, I tell them that I’m a problem solver. I use data to both find and solve problems.
My job is, firstly, to make sure all the hooks are in all the right places so that I can pull the numbers we need. Then, secondly, it’s to mash all the different sources of data together so they can speak to one another and, thirdly, it’s to massage the data so it gives us the insights we’re looking for.
While it’s important to sit and plan these stages, our data sources and back-end manipulations are growing so rapidly that we often don’t have time to organise everything perfectly. Our process is insights driven and as long as we’re getting helpful insights each week we’re happy.
Which metrics will matter most to a start-up depends on what stage of development they are at. At OpenAgent, we’re in a growth phase. So for us, some of the more important numbers to keep an eye on are the volumes of each of our sales stages. We need to ensure we are getting as many leads as we can from our marketing spend and then, that our sales team is converting as many as possible of these into tangible business.
To design metrics in such a way that it will drive behaviour in the way you want, it’s crucial to have a really good understanding of how the data maps onto the myriad of business processes which it reflects. You also need to sit with both the people who will be affected and the people who know the processes to understand their different perspectives. Then, you have to figure out how to get buy-in from the people whose behaviour you’re trying to change by incentivising them to change.
Knowing you are measuring the right things comes with equal measures of experience and good hard thinking. You know you’ve made it when you can isolate the exact cause of any change in any of the processes you report on. Until then, you’re not measuring all the right things.
But it doesn’t stop there. You should be reviewing whether your metrics are still relevant all the time. We’re moving so fast that we’re constantly creating new things that need to be measured, and the old metrics only stay relevant until the business processes change.
My biggest tips to other start-ups looking at designing good metrics is not to get hung up on outputs. Focus on the insights! That means that a lot of the time, “done” is better than perfect. You need to hone your ability to decide when you will be able to get your insights from a quick hack and when you will need to put some solid hours into coding a perfect data table. Don’t get hung up on trying to make something that’s failing work. Try everything, fail fast, learn and iterate.
Of course the good design of metrics requires having the right software and analytical tools.
At OpenAgent, we obviously use Excel, but we are also getting into using R (a programming language and software environment for statistical computing and graphics). While we have both MySQL and SQL Server databases, our greatest data asset is our business intelligence tool, Sisense. It’s the place where we mash every together in the back-end and produce clean dashboards on the front-end so that the different teams can monitor their work and progress. It’s like a really, really (really) fancy Excel, with lots of pretty pictures and features.
To succeed in designing metrics for a start-up, you also need a logical, creative and mathematical mindset. And laziness! I spend a lot of time thinking about how I can do things more efficiently with less effort. What separates start-ups from other places is the need for a willingness to learn, being comfortable with failures, and good people skills. We spend a lot of time trying to figure out never before solved problems with internal stakeholders.
Most important, though, is to create a good team vibe and to love what you do!