Every company has reached the understanding that data is important. But a few years ago, we realised that people didn’t always know where to go from there. They would come to us and ask, ‘We’ve got all this data, what can we do with it?’
Now that’s quite a big question and it’s one with a vast spectrum of answers. It depends on all sorts of things: your market, your strategy, your pain points, your users, your business model, your problems, the problems that lie beneath the problems, and the next step you’re looking to take.
Ultimately, to make the most of your data, you want to commercialise it. For the record, we’re not talking about selling user data to other companies here. This is about using the information you have to make better decisions - for your customers, in your operations and in your strategy.
So let’s break that down.
1) Drill down on your intention
Before you can figure out what data is helpful, you need to work out what question you need an answer to.
In our experience, this isn’t an instant process. It’s a matter of drilling down into your organisation to see where the pain points are. We often conduct intensive interviewing to discover the best and most tantalizing opportunities that could be best addressed with a data solution.
Once you have identified the opportunities, you can begin looking at the technology that is able to provide you with answers.
There are three main technological pots you can draw from here and you may want a combination to achieve what you’re after:
Data integration: This involves collecting and harmonizing previously siloed datasets (sometimes hundreds) and making it visible to different departments in a way that’s useful for them to make decisions.
Software: If you build a bespoke application, you can use data to automate business processes, doing the dull and repetitive procedural work so your people can focus their energy on the decisions that matter most.
Data Science: Using the magic of algorithms and analysis to give your people higher quality information that helps them excel at their jobs.
2) Identify the underserved
When you’re mapping out your needs and determining your intentions, it’s important to identify who in your organisation is currently underserved.
These are the people in your organisation who don’t currently have the resources they need to do part of their jobs efficiently. They might be pouring huge amounts of energy into getting the information they need to make crucial decisions and hitting their head against a wall.
Plugging these people into the data-fed infrastructure can make a world of difference to what they can achieve. This is certainly what we found when working with a well known performing arts company.
In essence, the client’s team didn’t have access to the assets they needed to set pricing efficiently. They were working everything out manually and, as a result, they couldn’t react to real time changes in demand.
The solution was relatively simple. It was a matter of sourcing relevant information from a variety of locations and synthesising it within an application. Our brief inevitably changed over time but soon the client’s team had the analytics platform they needed to offer a highly competitive service.
3) Don’t make decisions based on junk
There was a saying popular with early computer scientists: garbage in, garbage out.
It’s just as true with data science today. Cleaning data is the bugbear of pretty much everyone who works with data. If you have a phone number in place of an address, a location named ‘xyz’ or a ticket price that’s in the negative, it’s not going to make for very helpful analysis.
Cleaning data is essential but that doesn’t mean you can’t take a few sophisticated shortcuts. In an ideal world you would have completely clean data but if you’ve got some junk in there, you can set up some rules for your databases.
For instance, we built a data validation tool for a healthcare client that uses machine learning tools. It worked a little like a spelling/grammar tool such as Grammarly. But instead of suggesting ways you could rephrase an email, the tool would scan records for very specific medical ontology, and it would show you where you’d violated an assumption you’d made. Ultimately, it helped sift through the junk so better quality decisions could be made.
4) Opening the net up
Map out what you consider to be relevant data and then consider casting your net wider.
For instance, you can step beyond numbers and figures. Customer reviews can be integrated to reveal what people care about, what’s not working for them and what’s ripe for improvement. We did this for the University of Exeter’s Vista AR project, which used Augmented Reality to bring cultural and heritage sites to life.
As an alternative to combing manually through visitor reviews, our Natural Language Processing algorithms extracted that data that mattered: what was commented on and how the visitor felt. Key decision makers on the Vista project could simply check the dashboard we created to get the review analysis they needed.
If you’re struggling for ideas for how to use your data, look to other organisations for inspiration. It’s even worth exploring beyond your sector, since many problems are common to all organisations. See what they’re doing and what’s working, then borrow an analogy that maps onto your own industry.
5) Don’t build view-only business intelligence
Finally, you don’t want to build dashboards so you can just look at data. To get real value from your data, you need to make decisions from it. So any dashboard should show you what you need, help you diagnose it, and enable you to take action - preferably through the system.
If the place you view your data is the same place you act to address problems, you’re going to feel the difference in your efficiency. You’ll see tangible impacts in your business sooner, customers will reap the benefits faster, and your decision-making will be better than it’s ever been.
To learn more about how you can commercialise your data, don’t hesitate to get in touch.