Guest Post: Demystifying personalisation – above all, start small

Guest Post: Demystifying personalisation – above all, start small

Ah, the e-commerce buzzword, how we love to spout you – “single customer view”, “big data”, “attribution”, “programmatic”, “SoLoMo”.

By Chris Dalrymple, head of e-commerce at On the Beach

Ah, the e-commerce buzzword, how we love to spout you – “single customer view”, “big data”, “attribution”, “programmatic”, “SoLoMo”. These broad and ambiguous keywords are touted far and wide.

They fill some of us with dread, worried that we are not doing what we should, that our competitors are doing more than us, or, if we’re really honest, that we don’t completely understand what they mean.

But come they will, every few months a new one – in sales calls and conference agendas.

Take “personalisation” – one of those things that apparently everyone knows how to do but not many people seem to actually be doing. And we all should. Personalising the user experience is one of the most important things, probably second only to usability testing and split testing in the e-commerce playbook.

Keep it simple

It needn’t be complex.

Does your website remember what customers enter into the search form next time they arrive? Yes? You’re personalising. It’s a small thing, but make no mistake, it’s helping your conversion rate. No? Then you know what to ask your IT team when you next meet with them.

Do that before you invest in your retargeting campaign. It’s easier to do and probably more effective. Your customers will thank you.

Does your CRM campaign segment your customers based on recency, frequency or value? Who opens and who clicks? Yes? Good.

How about by their customer satisfaction scores? Couples, families and groups? Departure airports? Their last order details? Average order values?

Is your homepage different for customers you’ve seen before? Those based outside the UK? Can you simplify the experience if you know they just want to search? Or manage their booking?

Start simple and learn – take the data from your first test and do something with it. Test, learn, iterate. Keep going. Prove the value to the business.

We’re lucky in travel – there’s an implicit understanding from customers that they will give us a (comparatively) large amount of information and in return we show them holidays, flights or excursions. Use that information. Huge amounts of data, lots of repeat searches, lots of facets about what they booked last time. All of it points to what’s important to that customer. Make sure you can track it and do something with it.

If your customers are doing multiple searches, make this process as simple as possible. Let them hit the ground running on your website when they’re shopping around. Remember destinations, addresses, phone numbers and filters. Why would you make them enter it all again?

Personalisation doesn’t need to be about complex algorithms, data scientists and machine learning, at least not at the beginning. It’s about giving customers the softest possible landing when they hit your site and using the vast amounts of data you have to make their life easier and find what they’re looking for.

Run the data and look at the similarities between customers. How can you get the right product in front of this customer in the shortest amount of time?

Plug it in

Of course there are a host of third parties who can help us personalise and some of them are excellent. Many businesses don’t have the development resource to do it themselves. But buyer beware – the key is to understand how and why you’re improving the experience.

Start with the customer, come up with an idea, measure the outcome, understand what happened. Don’t let a line of JavaScript be a sticking plaster, and don’t let someone else learn more about your customers than you.

Definitely don’t set and forget – by doing that you lose a huge part of the value. Don’t underestimate the value of serendipitous discoveries.

Some of our best techniques to personalise CRM came from quick tests on our website. We would never had made those discoveries if we didn’t have an implicit understanding of how those early tests changed user behaviour.

Forget ‘big data’ – it’s just data. You can either do something with it, or it’s pointless. If you don’t know what happened today, how do you know what to do tomorrow?

How we do it

What about us at On the Beach? I work with a large, agile development team. A team of 70 developers with a permanent team focussed solely on personalisation – real time MI data with developers releasing to the website multiple times a day.

We’ve built a personalisation platform from the ground up over a number of years. But we started with a small project – a simple test – and went from there. Nothing complicated, not weeks of development, but we learnt something. And we’re still learning. I could give you 50 things we’d like to test next, and every time we run a test, the list grows a little.

Conclusion

If you haven’t started yet, you need to start soon. Develop a test and learn the approach. Look at your customer touch points, track how customers use your website and come up with an idea. Put something small live, see what the data tells you, improve the reporting if you don’t have visibility, learn from the results and do something else.

If you’re not already, sit with your development team – move desks. Ask them what’s possible. Give them the freedom to test quickly. Do five small things that might help you to learn about the ‘next big thing’ and understand the outcome.

This isn’t rocket science, and it’s more about business processes than algorithms. If your company is set up to develop quickly, to measure accurately and to come up with new ideas then you can succeed. If not, then find a way to get it there.

Above all, start small!