The five steps to data success

Natalie Cramp, CEO at Profusion

Data is playing an increasingly fundamental role in making and breaking businesses. Get it right and you can unlock insights and intelligence that can power your company to the next level. Get it wrong and, at best, you will lose out to your competitors, and at worst, you risk making ill-informed strategic decisions that will lead to real business problems. We hear a lot of data projects fail to deliver the value they should, so how can you approach your data journey to give your organisation the best chance of success?

When I launched Profusion’s Data Academy, our Data for Leaders course was really everything I wished I had known about data in previous organisations, which would have made me more effective, efficient and able to reach my strategic goals more quickly. Most of all, it probably would have enabled me to sleep a few more hours each week, which let’s be honest as leaders we all need. So here’s my simple five step guide:

Step 1) Start with the business problem - never the data

Most businesses have a lot of data. There’s usually even more that they could collect. All of this information can be analysed in myriad ways to find out all kinds of things. It’s a lot like playing Scrabble. The letters are data and the board is your company. There are a lot of ways you can put your letters together but only a few ways that will have meaning and maybe only one or two that will fit on the board to help you win. Good Scrabble players look at the board first to see where there are gaps and opportunities. Then they look at their letters to see how they can get the words they need. 

I cannot tell you how many organisations I meet where they tell me they have a data strategy and proudly give me a data architecture diagram. This is not a data strategy. Far too often people waste so much time and money getting all their data into shape, but only a tiny proportion of it is needed.

To get value from data, you need to forget the data and start with the basic problem you want to solve. It could be to increase revenue, reduce staff turnover, decide what product you need to develop next or make your marketing more efficient. The question you need to ask yourself is why are you using data? What problem are you trying to solve? The next step is to pose this problem as a question. 

Step 2: Get your question right - make it strategic, measurable, actionable and well-scoped

Let's say you want to increase revenue? The question could be, ‘what is preventing us from increasing sales?’. But let’s get more specific - we know some of the general factors at play in sluggish sales: lack of repeat customers, churn, low conversions etc.. So, maybe your initial question statement is, ‘I want to understand which of my customers churn and when, so that I can target them with win back campaigns and therefore increase customer loyalty and revenue.’

By articulating a clear question you want the answer to, we now know what data will be most relevant to collect and analyse. There is a clear scope. You also know that the insights that will be produced should answer that question - it makes the data project measurable. Any answer should also therefore be actionable. Your data tells you who is churning and the behaviour that is a key indicator that it’s going to happen - you can target them with a personalised reengagement marketing campaign, just before this does, and achieve your strategic objective of increasing revenue.

It sounds basic, but spend time focusing on asking a really good question. Time spent here will save you a fortune in time, effort and morale.

Step 3: Don't leave data people on their own - it's a team sport

Collaboration is king when it comes to a good data science project. Just like F1, it’s not just the driver who is responsible for winning or losing on the track, there are a whole host of people and different skills which combine to get you to that podium. Data is the same, there’s no point in hiring a couple of great data technicians and leaving them on their own. You need to ensure all skills be they technical skills, strategic understanding and commercial accumen, domain expertise and understanding of the end user, or organisational skills to ensure things run on time and to budget are all in the room. A lot of these skills exist within your organisation already, so it’s about ensuring you apply them to data projects, and don’t leave technical specialists coding in a corner.  

In addition to the specific teams you put around data projects, the more your data team knows about the business the better they will be at solving your problems. It’s also a two way street. Your wider team will naturally get more knowledgeable about data science and will then be better equipped to both interrogate the results and action the insights. 

Step 4: Prove value - start small before you scale

One of the biggest stumbling blocks to embarking on meaningful data projects is time and resources. It can be scary, especially when budgets are tightening, to spend money on a completely new intuitive. That’s why you should always start small to mitigate risk. At Profusion we like talking about POV (Proof Of Value) rather than the more commonly used POC (Proof Of Concept). What you should be proving is not that you can do it but it can create the value you think, before you make big investments or big promises to the organisation.

Remember the question above? A reason we want it to be specific is because a focused question requires a smaller project to answer it. It also requires less ‘buy-in’ and resources to action it. When you run your re-engagement campaign you can quickly measure the result to see the impact. This will provide clear ROI and make it much easier to justify ramping up your data initiatives. You could start with a much smaller cohort of customers and run some A/B tests, this will give you confidence before you switch off BAU communications to this cohort of customers and approach them in a more data-driven way.

Small projects also have the virtue of testing out your processes and infrastructure. You can identify problems that can be fixed quickly and easily before you scale up. 

Step 5: Technology is the icing - not the cake

So my final step is more a reminder. A reminder that despite what many #tech companies will tell you, that technology will only ever be the icing, not the cake itself. There are some amazing technologies available, and they can support our organisations fantastically. However, they will not solve our problems and they will not work if they underlying data that you have flowing through them and the business processes that you need to enable them are not there. They will not work if the people you are expecting to use them do not understand the insight they are meant to take action on, or how to use the system that is serving them. Technology companies promise you the world, it’s their job to sell you the dream. Technology enables things at scale, but it is rarely the solution in itself. Get clear on the business problem you are solving, and how you are approaching it, what insight you need and how you bring this together, before you layer technology over the top.

If you want to learn more about how to set your data transformation up for success then be in touch at dataacademy@profusion.com we would love to see you on our next Data for Leaders programmes or one of our many other programmes to support your organisation get the most value out of their data.