Case study · Experimentation & conversion optimisation
How rethinking a credit application improved online completion by 14%
A checkout bottleneck, a lengthy credit application and a deceptively simple question: could we make the process easier to complete without removing the information the business needed?
The problem wasn't getting people to the checkout
We had already spent time optimising the landing pages and product pages. Those parts of the journey had improved, but the checkout had largely been left alone. As the upstream experience improved, the application became the biggest remaining bottleneck.
There was a reason this part of the journey was challenging. It wasn't a conventional ecommerce checkout. Customers had to complete a substantial online credit application, including personal and financial information. They would then need to visit a store with supporting documents to finalise the process.
Improving the number of people reaching the application was of limited use if too many abandoned it before completing the online stage.
Understanding where the friction was
We used Google Analytics alongside heatmaps and click maps to investigate how customers were interacting with the application.
The evidence suggested two related problems. Some customers appeared to encounter the full length of the form and leave before they had really started. Others progressed partway through before abandoning, potentially because the process felt too long or asked for information they didn't have to hand.
That included things such as bank account information, National Insurance details and identity documents. These were plausible explanations for abandonment, rather than reasons we could definitively assign to individual customers from behavioural data alone.
The hypothesis
If we reduced the perceived complexity of the application, showed customers how far they'd progressed and made the time and information requirements clearer upfront, more customers would complete the online application.
Rather than testing five isolated ideas, we designed a combined intervention to address the friction identified in the investigation.
What we changed
One step at a time
We divided the long application into steps. Subsequent sections remained on the page but were hidden until customers reached them.
A clear step indicator
Customers could see their position in the application instead of facing one uninterrupted form.
A realistic time estimate
We communicated that the application could be completed in less than four minutes. The message was informed by a measured average completion time of approximately three and a half minutes.
Requirements upfront
The opening paragraph made clear that customers would need bank account information and identity documents, reducing the chance of encountering those requirements unexpectedly.
Fewer competing routes
We enclosed the application experience by removing extraneous navigation, helping customers focus on the task they had started.
An important distinction: online completion wasn't the end of the journey
The experiment focused on completing the online application and credit-check stage. Customers still needed to attend a store to provide supporting documents and finalise the process.
That distinction matters because an uplift in online application completion should not be described as an equivalent uplift in completed sales or credit agreements.
The result: from test uplift to reported commercial value
The redesigned application produced a 14% relative uplift in online application completion during a four-week A/B experiment.
The initial annualised commercial estimate was approximately £25m. The figure subsequently attributed to the improvement in year-end reporting was approximately £20m over the following year.
Those are different measures. The first is an experimental change in the online application completion rate. The second is a projection. The third is a retrospective reported commercial outcome, dependent on the wider customer journey and business performance.
Because the intervention combined several changes, the test does not tell us which individual element drove the result. It supports the combined redesign, not a claim that the progress bar, step layout or navigation removal alone delivered the uplift.
What this taught us
Long forms aren't always a problem simply because they're long. Sometimes the problem is the amount of effort customers think they will have to invest, the uncertainty about how much remains or the discovery that they need information they weren't expecting to provide.
In this case, the business still required the same application information. We focused on making that requirement easier to understand and work through.
There's also a broader lesson about prioritisation. Once the landing pages and product pages had improved, the biggest opportunity had moved further down the funnel. Continuing to optimise the same parts of the journey would have missed the more significant constraint.
A note on evidence and attribution
This is an anonymised retrospective case study of work undertaken more than a decade ago. The original experiment dataset and detailed statistical outputs are no longer available, so sample sizes, confidence intervals and other unverified statistics have intentionally been omitted.
The four-week experiment, 14% relative uplift, approximately £25m projection and approximately £20m year-end reported value are presented from the consultant's recollection of the engagement. The historical year-end document has not been independently verified for this publication.
The experiment measured online application completion, not final in-store verification or the overall suitability of credit decisions. The commercial outcome should not be interpreted as a directly observed four-week test revenue uplift or as proof of improved customer outcomes beyond the measured application stage.
