Let’s assume you’ve been using Google Analytics for some time in your organisation and your digital marketing endeavours at least are data-driven. More and more people are starting to see the benefit in using data to get their jobs done and now everyone’s talking about Data Science or Big Data or whatever. Let’s be clear – when we say “data” we are not talking about just Google Analytics but the proliferation of this tool means that it is very likely a starting point for your data analytics. Alongside your Google Analytics data, you’ve probably got a CRM and a bunch of other data sources that you know you should be doing something with – and gaining some kind of “competitive advantage”.
Better get a plan! If you don’t you’ll be left behind, right? Well – yes and no.
You do need to know where you are going. Some form of stated intent, whether that be rough notes or a multi-stakeholder workshop-derived several-chapters-long document, will be needed to keep you in check. Start with the “why” then the “what”. Why do you want to use data? What data do you need? Be open to yourself about these questions – you might not know; and that’s fine!
But stuff changes, fast. You can’t plan for every eventuality. And your organisation’s stance will determine the roadmap for your data strategy.
According to a couple of clever chaps writing at Harvard Business Review, there are 2 types of data strategy that every organisation needs – an offensive and a defensive data strategy; where an offensive data strategy is focused on flexibility and a defensive data strategy is focused on control. You will need a mix of both and that mix will depend on your industry sector, competitive landscape and current trends.
While such a model is useful to understand your organisation’s stance, it doesn’t have to be laboured over. In fact, acknowledging where you currently sit in the mix and understanding this changes over time is enough. As your organisation progresses along their path of analytics maturity, your stance with regard to data strategy is likely to alter too.
To get things started, let’s take a key principle from the Lean Analytics movement – just “measure what matters”. Take a look at your business model and ask yourself – do you know your customers? Do you understand their buying process? Do you know why and how they buy?
Once you’ve examined your business model to determine your goals, you will need to assess the data you have available and what data you need to start collecting (we do that here with what we call a Measurement Framework). There might be a trade off; ask yourself what’s the value you will likely derive from all this work? Some technical implementations can be costly and that cost simply might not measure up against the returns it could bring.
Some plans might entail investing in new IT infrastructures, such as BigQuery for Data Warehousing, or new hires, such as taking on an in-house analyst. But before you do, ask yourself what can you do with external help? And you will be pleasantly surprised quite what can be done with your existing systems at very little comparative technical overhead these days. New ways to pipeline and manipulate data are emerging all the time.
Most importantly, can you trust your data? It is a rare thing to hear of a company that has an absolutely fault free end-to-end data collection and analysis pipeline. And most of the issues start right at the very beginning of that pipeline – with a misplaced tracking tag in Google Tag Manager, for example.
To help ourselves in this process, we have created the “Data Strategy Canvas” (nod to Business Model Canvas).
It is still in early stages (version 0.1 in fact) but those of you familiar with the canvas approach will quickly grasp what the intention is here. Draw up the canvas onto a whiteboard or print it out on a big sheet of paper, then simply use post-its to capture answers in the corresponding areas.
So, before you launch head first into your Data Strategy, perhaps you might find this canvas of some use 🙂
Is your analytics data ship-shape and tickety boo? Can you make business changing decisions on the back of the data you have? And, before you start hiring a team of analysts, can your existing team be trained to a reasonable level?