Jono Alderson describes himself as a closet web developer, turned Technical SEO and analytics geek, and now focuses more on data analytics strategy. The key thing he points out in this talk is that data itself isn’t actually useful at all – it’ frameworks for understanding that data which will drive success.
We don’t do enough work to understand what our success looks like, and so when we get lost in data and cohorts and APIs, it turns out we’re not actually generating much value for our organisation. What’s more common is that, under pressure from management, we start digging through our data hoping to find some magical nugget of insight, rather than defining clear goals for our data. But that’s a trap.
The first problem is that we generally have difficulty articulating what “good” looks like for our products. The second problem is that, as a result, we end up simply framing our goals relative to past performance, or our competitors, rather than a more objective view of success.
The Definition of Success Varies
Ultimately, the key to getting over these challenges is to recognise that you have bespoke challenges, and therefore you need bespoke data. Or rather, a framework for understanding the data you have access to. To that end, Jono provides a fine-grained, detailed structure to describe exactly what success looks like for your products.
The first thing you need to do is get all of your key stakeholders into a room, and get coherent agreement on what success is. Getting consensus will be harder than you think, but will allow you to start a more constructive conversation. To start tracking your new, clearly defined objectives, you will almost certainly need to invest in new infrastructure, and acknowledging this at the start of this progress sets clear expectations.
The next step to effective data analytics is self-improvement – having a system for assessing where you’re strong and where you still have to improve your data collection and analysis processes. As a general rule of thumb, judging and acknowledging your data maturity should come first, and tools should come later. As you improve your understanding of exactly what you need to be tracking, you should be building your analysis criteria and your KPIs directly into your product specs – i.e. make sure you know how to decide if a new feature is successful or not before you launch it.
- Make sure you know what success or failure looks like before you start exploring data.
- Get your stakeholds locked into clear definitions of success.
- Have a clear process for self-assessment so that you’re always aware of how mature your data collection and analysis processes are.
Ultimately, this should help you to make decisions about things that matter!