According to a recent Gartner report, only between 15% and 20% of data science projects get completed. Of those projects that did complete, CEOs say that only about 8% of them generate value. If these figures are accurate, then this would amount to an astonishing 2% success rate.[1]

According to these statistics, most data science projects are never completed. Of those that are completed, only a fraction prove valuable.

This creates a worrying situation for data science teams. And raises the question of how long the leadership groups within their organisations can support expensive data science departments that appear to produce little value?

The same can be said of marketing analytics. Gartner reported last year that over half of CMOs are dissatisfied with the impact of their marketing analytics teams. As a result, Gartner predicts that marketing analytics teams will lose headcount – with many teams being slashed in half – within the next three years.

In my experience of marketing analytics, however, there is an important distinction to be made between perceived value and actual value when judging the impact of an analytics or data science team.

Unfortunately, the fate of data science and analytics teams can rest more on perceived value than actual value.

‍Actual Value vs Perceived Value

Actual value is the true impact that marketing analytics is making.

Perceived value measures the extent to which people in the business recognise that impact.

Actual value requires that the analysts or data scientists are good at their jobs, work on the right problems, deliver outputs that the organisation uses and that make a positive impact on KPIs. To a large extent, actual value is within the control of the data science and analytics teams.

Perceived value requires that people are aware of the actual value and, importantly, when asked, they remember it and will attest to the value analytics and data science are delivering. As such, this requires that data science teams make a strong positive impression on the right people, and regularly provide them with evidence of their data science successes.

Closing the 'Value Perception Gap'

I worked in a very large company once, with hundreds of analysts and data scientists in the analytics department. We were growing the team rapidly and wanted to make sure the business understood the value that it was getting for its investment in data and analytics. This was important for us, as we wanted to keep growing and expanding our reach within the organisation.

We introduced a benefits tracker. This simple idea proved hugely powerful and gave our team a compelling way to close the gap between our actual value and our perceived value.

Within months, the change in status we were afforded was palpable. The biggest difference? Quite simply, the business started to be aware of the value we were delivering.

That’s why I believe every data science team should maintain and promote its own benefits tracker.

‍The Analytics & Data Science Benefits Tracker

A benefits tracker is a simple spreadsheet. Here’s a template you can view and download for free.

As you'll see, it's a way of tracking all active and completed data science and analytics projects, showing the name of each project, when it completed and the value that it generated. You may wish to track more or less information about each project.

The important thing is that every analyst, manager and data scientist updates the benefits tracker on a regular basis.

The result is a single spreadsheet that captures the value – in black and white – that data science and analytics delivers to the organisation.  

Why You Should Start a Benefits Tracker Today

The advantages of the benefits tracker are significant.

But some are more obvious than others.

Here are nine things that we’ve found data science and analytics teams gain when they start tracking their outcomes in a shared, single location:

1. Headline results. The benefits tracker gives the leadership team an easy way to see hard, aggregated measures of value that the data science and analytics teams are delivering. The question ‘are we getting value from data science?’ should never be answered subjectively. The benefits tracker gives an objective answer.
2. Cheerleading script. Everyone loves a simple message. And the benefits tracker makes it easy to communicate the value of data and analytics projects in big picture, commercial terms. For example, if the combined value of data projects is £50,000 last month (or £5m last quarter) – these types of big numbers make an easy, consistent and memorable message that everyone in the team can get excited about (and shout from the rooftops!).
3. Commercial focus. One criticism of data science teams is that they can appear overly academic, not focused enough on the commercial aspects of the business. By implementing a benefits tracker, this sends a powerful signal to the entire data science team that its work must have positive commercial outcomes. It also means that KPIs must be defined at the start of each project – one aspect of best practice that is overlooked surprisingly often.
4. Collaboration around results. The numbers you enter into the benefits tracker must be accurate, and therefore data scientists and analysts have to engage with their stakeholders in the wider business to understand and validate the true impact of their work. This type of following-up and interrogating the true value of a project is best practice, and the benefits tracker forces it to happen.
5. ‍Support for further investment. Data science teams – like any other – want to thrive and grow. The case for further investment (for example, in more people or new tools) is far easier to make when it can prove the value it’s already delivering.
6. Protection from funding cuts. The inverse of point five (above) is also true. Tangible measures of commercial value provide strong protection for the data science team should the business fall on hard times.
7. ‍Snowball effect. Once the organisation truly understands the benefits the data science team can deliver, demand for its input will grow. Whether it’s being invited into more projects, or different projects, the data science team will start to get involved in new areas and begin to deliver even more value as a result.
8. Saying ‘no’. Often projects can be commissioned because they seem like a good idea. However, if the KPIs are not clear at the outset, or the anticipated value is small, the process of completing the benefits tracker gives the analytics team a good reason to say ‘no’, or at least to have a more informed conversation about the risks with their stakeholders.
9. ‍Low value projects. Where projects don’t deliver the desired level of value to the business, the benefits tracker makes this obvious, and gives the data leadership group the opportunity to dig deeper. Was it the wrong project? Did something go wrong? Is there a pattern here?

‍

Advice for Setting Up a Benefits Tracker

The case for setting up a benefits tracker is compelling. There are, however, a couple of important points to note.

First, attribute successes accurately

Determining the actual value of analytics and data science projects can be trickier than it seems.

For example, getting agreement that it was your team that drove 50% of a 4% increase in sales last month may be difficult if the sales team also implemented a new initiative, went on training and brought in two new starts.

To attribute success to analytics correctly, it's best to get agreement from other departments (including Finance) or at the very least to be appropriately caveated in the comments.

Preserving the integrity of the benefits tracker is critical for it to do its job properly.

Second, measure 'softer' benefits as well

Not all projects will have hard financial benefits.

Of course, additional sales revenue and cost savings will form much of the tracker’s headline figures. However, some projects will have benefits that don't easily translate into monetary value. For example, projects may be required because they build capability, serve a regulatory requirement, improve customer experience or improve employee engagement. It’s just as important to keep track of these benefits via the correct KPIs.

‍Closing Thoughts

The concept of tracking value in such a black and white manner may be new – and possibly even uncomfortable – territory for many data science and analytics teams. However, if you feel there is a gap between the actual value you deliver and your perceived value within the organisation, a benefits tracker can be an invaluable tool in getting your department the recognition it deserves.

Also, the data science team that doesn’t evidence its ongoing value in commercial terms risks becoming isolated from the wider business. And keep in mind that senior leaders typically only ask for evidence of the value when there’s a problem. Much better to be on the front foot, and give your entire department a consistent, easy narrative that demonstrates your value.

Finally, as far as processes go, implementing a benefits tracker is usually straightforward. And, as I’ve hopefully got across, keeping track of the value you bring to the business is only one of its advantages. Done right, it will make your team more commercially focused, more commercially accountable and better able to improve your actual value.

‍

[1] Source: https://fastdatascience.com/why-do-data-science-projects-fail/