It is pretty much a given that great analytics can drive better decision making, inform product development and help identify where cost savings can be made. How to get there is the challenge!
Sadly, many projects will falsely identify fixing all internal data as the first task and fail (think Einstein’s definition of insanity, doing the same thing and expecting a different result; you have definitely tried to fix your data before!). Here are some thoughts on what you could do differently to create a data strategy for your business:
Don’t groan, thinking about your Key Performance Indicators is a useful exercise! Let’s use lead generation as an example. With the race to drive web presence, is your metric how many people visit the website (quantity) or is it concerned with bounce rate (quality)? While this is simplistic, it’s a starting point!
Okay, so what if I am really interested in website visitors that came from a specific route (let’s say LinkedIn) and visited three pages before going on to buy more than £1,000 of goods - by measuring this, I can understand if my LinkedIn presence is working for me.
The data itself is provided by website analytic engines and your sales order processing application. So you have the data, brilliant! Now you need to unlock that data and present it in a way that is useful to you.
Having a data strategy helps you focus on what data is important. Now you start to understand what data needs to be reliable and stop stressing over the impossible task of keeping all of your data squeaky clean all of the time.
We made an assumption above, we assumed that we’re working with data we trust! Cleaning data is like clearing a loft, something you try not to think about and leave until forced by a big event (new system implementation or house move). Let’s extend the house move analogy and ask yourself this question: Last time you moved house, did you find an unopened box from when you moved in?
Part of your data strategy needs to include housekeeping. If historical data is important then shape it up. If less so, then don’t sweat it. Either way, put energy into managing data going forward. Your existing applications should be helping with this. Minimise free text fields, use mandatory fields, workflows/approvals and make use of de-dupe tools that all help ensure your data smells sweet but only do it on the data you need to analyse (and where possible automate the capture), otherwise you will make your processes laborious and as water runs downhill, your users will look for ways to side-step.
Data analytics (and big data) are nothing new. Back in the 1950s businesses were using basic analytics, typically manually reviewed spreadsheets, to uncover insights and trends. It’s common sense that if you capture all the data that streams in your businesses, you can apply analytics and get significant value from it. So why isn’t it common practice?
The new benefits that big data analytics bring are speed, efficiency and accessibility. Whereas a few years ago a business would have spent considerable time/money to gather information, run analytics and identify those trends, today it is easier than ever to identify insights for immediate decisions.
The ability to work faster is what generates the ROI; there’s no return on data that tells you what you should have done, you want to know what to do next.
And whilst we’re on measuring, managing and ROI, think carefully about monitoring the cost. “Hot path” data is expensive to keep so consider what needs to be hot and what could be cold. Tools such as Microsoft Azure Data Services help you structure the storage and refresh options.
Three years ago, a McKinsey Global Survey showed that business leaders expected their analytics activities to have a positive impact on company revenues, margins, and organisational efficiency. A revisit revealed mixed success: issues with strategy and tools were not the biggest culprits. As with any project, the key to success is in the tactics and people: leadership support, communication and right resource for the job.
The right resource is not always about the previous experience; the tools are easy enough to use with the right mindset and support to get to grips with them. You can (and dare I say, should!) “grow your own” resource and if you are engaging a specialist, remember the most successful projects are collaborations where the partner is developing your team (doing it “with you” not “to you”).
Once you have your homegrown data champions, your team will refine the data strategy. This, in turn, will lead to a further reduction in the manual tasks associated with data preparation and management, leaving more time to focus on higher value activities.
Dolphin versus whale, don’t eat the elephant etc. Before you try to do everything all at once, identify a manageable subset. Use the tools to spin it up, monitor cost and value and refine before scaling upwards and outwards.
Data storage is relatively cheap, working with it is more expensive; but with the volumes of data created it is a cost that can soon escalate. Managing the cost is where a good partner will add value. I do predict some “bill shock” stories for those who don’t provision their data storage effectively.
One of the really great things about how rapidly these solutions can be created is that you can run inexpensive proof of concepts to quickly check that your harnessed data is moving you in the right direction.
Discover how you can get started today by downloading our Predictive analytics facts sheet.