Grass parallel lines_crop

OPINION19 July 2018

The making of a metric

Data analytics Opinion UK

Ryan Howard lays out three steps to building a metric that is worth tracking.

Unless you are an enthusiast of parallel and ever so slightly squiggly lines, research does not get any more tedious than a tracking survey. If not occasionally reimagined, it may tick over, year in, year out, achieving little else than boring a hapless audience to tears. Its original intentions, along with those thought to have designed it, forever lost to the sands of time.

Given that trackers command so much investment and commitment, and influence everything from strategy to remuneration, we ought to take great pains to ensure we are tracking the right things in the right way.

Brand trackers gauge brand health or brand trust. Customer experience trackers tell us about engagement. What are these but fuzzy, abstract concepts, which your colleagues may struggle to describe, let alone agree how to measure? Nevertheless, should customers be able to express these ideas clearly, we gain a metric that holds great sway.

Here are the three steps to building a data-driven metric that matters.

Step 1: Decide what Return on Investment (ROI) means to you

Building a brand by driving spontaneous awareness, positive word of mouth or improving customer experience are not ends in themselves. If we zoom out, we realise that we are far more concerned in the upshots of these activities; like decreasing customer churn, upselling or growing market share. Seek out the point closest to the bottom line where success can be recognised and quantified.

Step 2: Demonstrate a link

Pilot questions. Reserve time to study the relationship among and alongside your chosen ROIs. Here, it pays to move cautiously. If a survey question fails to correlate, bin it. The goal is to land on a questionnaire, which has the highest correlation with Step 1. This is the silver bullet. It is the reason to fund initiatives that drive the metric up and the reason to cut those that do not.

You may want to move beyond a single recommendation question, landing on a combination of measures, which work together in concert. The resulting metric will be more predictive and sensitive to the market dynamics that you deem relevant. A perspective devised indirectly by your own customers around ideas central to their understanding of your brand, rather than by committee.

Step 3: Shift it 

The final piece of the puzzle is to identify the drivers of your new metric. Focus only on decisions that could be taken in the next few years. Prioritise the pragmatic. Look to the low hanging fruit. These may be reducing queuing time, investing in staff or reconfiguring media spend. The drivers will change over time because your efforts are targeted at the very metric you are tracking. As you improve the metric, other priorities will emerge, like a never-ending game of whack-a-mole.

As you play, a once dreary survey grows to be the most trustworthy predictor of future performance and a trove of diagnostics. The reason to believe is baked in from the start, which means not having to wait for two years of trending.

From the very first day, the language of commercial success is the same rally cry from boardroom to coalface. A great marketing campaign is only a winner if it shifts the metric, not because it is liked or particularly memorable. Similarly, customer experience is only worth investing in because it is profitable, not because it makes people think happy thoughts. If this data-driven approach is the starting point, your research, just like your squiggly lines, will be going places.