We experiment and test

Quantitative approaches don’t have to be about synapse-wilting surveys stuffed with questions. 

The best quantitative approaches recreate the context which you might encounter in real life, getting people to make real choices — often under time pressure. 

Such methods might include monadic testing, behavioural experiments and A/B testing.

How we do it: The £100m letter

Pension providers lose contact with thousands of their customers every year. The scale of the problem is vast: approximately 1.6 million pension pots worth £19.4bn are unclaimed. This has serious consequences for people when they retire. 

For the most part, the causes behind this are mundane: people move home, get divorced, or simply forget they have a policy. The industry tries to trace customers, but reconnection rates are low — with less than half of people responding. Sometimes, reconnection letters look too good to be true, so customers think they’re scams. 

That’s why the Association of British Insurers asked us to explore the behavioural ‘nudges’ their member firms could use to encourage lost customers to respond. So we created a controlled experiment to isolate the behavioural drivers of response, comparing an anonymised provider letter as our control with four new test versions.

Our redesigned letters were significantly more likely to get people to act. The target set for the project was a 10% increase in the reconnection rate, equating to a £100 million increase in assets being reunited with customers annually. 

Our Communications Guidelines were launched by a Government Minister and received extensive media coverage, including Martin Lewis raising the issue on ITV’s This Morning.

Coming up…

In our next article, we’ll be exploring how ‘triangulation’ can improve research projects.

Free copies of our book ‘Closing the Say-Do Gap’ are available now. Drop us a line to get yours today.