Many times when you launch a data project (or even when you’re pitching one) there are projected savings bandied around in the millions. Have you ever wondered where data value metrics come from?
Plucking numbers from thin air
Often, when looking at the project success of a data project we might fall back on gut instinct. Or some might do paper-based exercises that create rough ‘guesstimates’ that aren’t then validated. But data doesn’t work like that. It’s something that we recognised early-on. That’s why Cynozure doesn’t propose or advise on arbitrary numbers to our clients. Instead, we take a deeper look at the data and the organisation’s goals. I’m going to take you through how we do this so that you can truly unlock the value in your data.
A more robust approach
Instead, there’s a more scientific approach that you can take. It requires constant testing, learning and refining. The data value metrics you use will vary from other organisations (as well as industries). It’s vital to choose the right metrics in order to uncover the areas of your business where value can be added through data. But to start with, (somewhat misleadingly) to set your data value metrics, you don’t start by looking at metrics at all.
Cynozure’s framework for defining success
To illustrate this better, here are the steps run through to define success, and how data value metrics fit into this framework:
- Understand your goals
When looking at your goals, it’s important to differentiate between your personal goals (and those of your peers) and the wider organisation goals. To add value to your organisation, it’s vital not to prioritise goals based on those that promote and shout the loudest. Instead, consider the ones that will have tangible benefits to your bottom line – informing new products, reducing costs, and improving marketing effectiveness for example. Preferably in a short time frame so that you can prove value quickly.
- Uncover the ‘art of the possible’
The key is to get a wide list of different use cases and then narrow those down at a later stage. Take off the shackles of what has previously been possible and the restrictions you’ve had in the past. Forget if you tried something before and it didn’t work because that may not be the case if you tried again. Look to others in your industry for inspiration – or better yet – look at other industries to see if there is anything they are doing that could be reflected in your organisation. For example, could a local council learn anything from how digital retailers engage with customers?
- Identify the benefits
From your possible use cases, identify the benefits that each brings to your organisation. See where those benefits align with your business goals. Although some of these benefits are monetary, don’t forget about other ones as well. Improved data literacy and more informed decision making are two examples of this.
- Then to the metrics…
Once you’ve listed out the benefits (and chosen your use cases based on these) you can then get to your metrics. What do you need to monitor in order to understand whether your projects are meeting your goals? It’s important at this stage, amongst other things, to agree on the definitions of your metrics, the business rules required to calculate them and who the owner of the metric is.
- Test, test and test again
Come up with targets for the metrics and constantly test and learn from the results. Run A/B tests and see what gives you improved returns. At an early stage, it’s a good idea to have a lot of smaller trials that you can keep tweaking.
- Extrapolate positive returns
Once you have a few successes from your early projects, roll them out across your organisation. From those results, extrapolate (sensibly) to come up with the uplift value and then use these metrics for your top-level reporting. You are then able to put a value on what your data strategy is able to deliver
The framework in action
A leading theatre company approached Cynozure with the goal of improving revenue from ticket sales. In order to achieve this, Cynozure identified that the company had no single customer view (which was amplified by a lack of visibility on sales through third-party vendors). It also didn’t understand whether it was putting on the right shows in the right areas or seasonal variations in ticket sales.
To start, Cynozure ran a series of trials in different markets and areas. Data was collected from market research, social media, and online transactions. The trials ran over a three month period and helped the theatre company to better understand customer purchasing behaviour. The insights gathered from these trials then informed a new data strategy – plus the metrics that the company could use. These included ticket sales uplift and how many seats were sold each night.
The company is now rolling out its data strategy and a related data analytics platform that surfaces key insights across the organisation.
An essential step
Without the right metrics, you’ll never truly understand whether your data projects are delivering value or not. Starting with your business goals, then choosing your use cases and metrics is the best way to demonstrate value to your organisation. It’s a crucial component of every data project. Often, it can be tempting to dive straight into the data with no way of measuring its effectiveness. But that’s a short-sighted approach that will ultimately undermine your efforts.
Jason Foster – Founder & CEO