As an industry, we are in the early days of personalisation. Technology has come a long way, and there is likely still a long way to go.
How much we invest now and in what remain big questions for media organisations.
Is it worth putting a lot of investment into active personalisation for our products if only a percentage will take it up? Should you build your own or is the technology ready — or soon to be ready — for out-the-box use?
Below are a few other remaining questions we don’t have perfect answers for:
1. How do we measure personalisation?
Currently, tests are measured by click-through-rates (CTR) as that is the most instant, easily comparable metric.
But we know CTR is not the best business metric overall as we often want engagement and retention. So how do we look at the more lagging metrics?
A/B tests need to happen, but longer-term retention is hard to accurately predict. And there are so many other inputs that can affect the results, which makes it nigh on impossible to identify which markers had what effect.
NZZ in Switzerland noticed an uptick in completion rate, suggesting articles shown are more relevant. Perhaps we can also look at session lengths.
2. How do we communicate to readers?
There is a level of hesitancy in telling users that we are personalising their content for them based on their behaviours and other inputs. It may even feel a little creepy to some. But most users are now used to it. Many expect it, especially those brought up with digital content streaming services.
3. What is the right level of transparency with users to retain trust?
Clearly with active personalisation, this isn’t a question. But with passive, we need to balance explainers with making them too aware/worried about it. It makes sense to have an exampliner accessible that allows curious readers to learn more but doesn’t showcase it too much as a huge change they should be worried about. Here is an example explainer from The New York Times.
4. To what extent do we balance user needs with business needs?
Most companies are just getting started and, rightly, focusing around user needs. We’re starting to see some optimisation for business goals such as Schibsted’s in-house algorithm and The Globe and Mail’s Sophi maximising to convert subscriptions. And there is a lot further we can go.
Once people have registered, can we optimise to build daily habit? For those who are unlikely to subscribe, can we optimise for advertising? Perhaps this can be even more nuanced, prioritising more profitable formats such as video where we can.
As these algorithms get more sophisticated and we learn more about our audiences, there is much untapped potential.
If you’d like to subscribe to my bi-weekly newsletter, INMA members can do so here.