How Harborline Brokerage cut account-opening drop-off by rebuilding verification

Harborline Brokerage is a fictional retail broker used throughout this site for illustration. The figures below are invented to show how a case study is structured and do not describe any real firm.

The problem

Harborline's marketing brought a steady flow of applicants to its account-opening page, but only a little over half of those who started an application finished it. Analytics showed the largest single drop at the identity verification step, where applicants were asked to photograph a document and take a selfie. A further drop came at the funding step, where applicants who had passed verification were asked to link a bank account by entering details manually.

Compliance rules required identity verification before an account could be funded or traded. The firm's fraud team was reluctant to loosen the check, and manual review of failed verifications was consuming most of the onboarding team's time.

What they did

The firm started by measuring rather than redesigning. It instrumented each step of the flow, tagged the reason for every verification failure and interviewed forty applicants who had abandoned. Three findings shaped the redesign.

First, most verification failures were image-quality problems, not identity problems: glare on a document, a cropped edge, a selfie in poor light. Second, applicants who failed once and were asked to try again rarely did; they left. Third, the funding step was asking for bank details that most applicants did not have to hand.

The redesigned flow moved verification later, after the applicant had chosen an account type and answered the suitability questions, so that they had invested effort before the hardest step. The document capture was rebuilt with live guidance on framing and lighting, which reduced first-attempt failures. Applicants who failed the automated check were offered an immediate alternative, a short video call with an agent, rather than a retry.

Funding was switched to an open-banking connection that let applicants authorise a transfer from their banking app without typing account numbers, with manual entry kept as a fallback.

The fraud team reviewed each change against its own metrics and set a condition: any increase in fraud losses at onboarding would reverse the change. It also added a second-line check on accounts opened via the video-call path.

Results

Over the six months after launch, the share of applicants who completed an application rose by 31 per cent relative to the prior period. The median time from starting an application to holding a funded account fell from just over two days to about six hours, mostly because of the funding change.

The manual review rate on verifications fell from 18 per cent to 7 per cent, freeing the onboarding team to focus on the video-call path. Fraud losses attributable to onboarding were unchanged as a share of new accounts, which satisfied the condition the fraud team had set.

What others can take from it

Measure before redesigning. The intuition inside the firm had been that the verification requirement itself was the problem; the data showed it was the execution.

Do not make the hardest step the first one. Applicants who have already made choices are more willing to complete a demanding check.

Replace retries with alternatives. A second attempt at a failed step is where most drop-off happens; a different path keeps the applicant moving.

Give the control function a veto with a measurable trigger. The fraud team's condition made the change acceptable to them and kept everyone honest about the trade-off.