Backstory

Customers were showing up ready to switch, 76% of visitors who landed on the funnel started a broadband quote. But that intent collapsed almost immediately afterward only around 18% of people who completed a quote went on to click through to an actual deal, and just 15% of those clicks turned into a completed switch. Because the deal results page sits entirely within ApTap's own design and content, it represented the clearest, highest-leverage opportunity to close the gap between intent and action, without needing to touch upstream marketing or downstream banking integrations.

As Product Designer, I was responsible for diagnosing why the drop-off was happening and designing a solution that could be validated through experimentation before being handed to engineering.

The starting point was a clear discipline: the funnel data told us where customers were dropping off, not why. Rather than jumping to a redesign, I treated the initial numbers as evidence of a problem, not a diagnosis, and built the project around a structured narrative; Data -> Questions -> Hypotheses -> Research -> Insights -> Design -> Experiment -> Outcome.

I framed the central design question as a How Might We: How might we help customers quickly understand and confidently choose the broadband deal that is right for them? 

Measuring success

The recommended path was a controlled A/B test. The existing deal results page as control against the guided recommendation, transparent comparison and reassurance content as the variant, measuring the primary deal-selection metric while monitoring switch completion as a guardrail, so any lift could be trusted as real improvement rather than a shifted bottleneck. 

The Problem

Research plan

The Process

Six candidate explanations for the drop-off shaped what the research needed to test

• Choice overload; too many similar-looking deals making the decision feel harder rather than easier.

• Poor differentiation; customers unable to tell what practically separates one package from another.

• Weak value communication; features listed without explaining which option is actually the best fit.

• Price uncertainty; unclear promotional pricing, post-promotion pricing, or contract terms.

• Lack of reassurance; unanswered worries about installation, cancellation, timing or hidden costs.

• Weak next action; a primary CTA that doesn't make the next step obvious.

Reviewing the existing deal results experience against the funnel data surfaced a working problem statement; customers who complete a quote are not confidently progressing to a deal selection, because the results page may not be doing enough to help them understand their options, evaluate value, resolve uncertainty, and feel ready to commit.

I mapped this against the full customer journey; Discover -> Start quote -> Provide information -> Complete quote -> Review deals -> Compare options -> Decide -> Start switch -> Complete switch, and identified the deal results page as a key psychological transition point: this is where the customer shifts from simply providing information to making a financial and contractual decision. That reframing mattered, because it meant the page needed to support confidence, not just present information.

ApTap connects banks and merchants to help customers save.

Initial UX hypotheses

Recommended deal card; Each recommendation surfaces the deal name, headline speed, monthly price and contract length alongside a plain-language "Great for" description, a short "why we recommend it" explanation, key terms and conditions, and a prominent "Choose this deal" CTA - so the decision-relevant information sits together instead of being scattered across the page.

Solution exploration

To move from hypothesis to evidence, I scoped a four-part research plan:

• Funnel and behavioural analytics; reviewing abandonment at each step, measuring time on page, scroll depth, deal views, expansions, CTA interactions and back navigation, and tracking the full event sequence from quote_completed through to switch_completed.

• Usability testing; recruiting people who had recently considered switching broadband providers, giving them a realistic task (review the available deals and choose one), and observing comprehension, comparison behaviour, hesitation and decision confidence.

• Heuristic review; evaluating clarity, hierarchy, cognitive load, transparency, feedback, accessibility and CTA visibility against best practice.

• Competitive review; studying broadband and adjacent subscription journeys for patterns in recommendation, comparison and pricing transparency, used to inform interaction patterns rather than visual style.

Validating the direction

The primary metric was the quote-completed -> deal-selected rate, against an indicative baseline of 18%. Secondary and guardrail metrics; deal-selected -> switch-started, switch-started -> switch-completed, results-page abandonment, time to deal selection, number of deals viewed, and support/cancellation rates, were tracked deliberately so that a lift in deal clicks couldn't be mistaken for success if it came at the cost of downstream switch completions. 

Key learnings

The most important discipline on this project was resisting the pull to jump straight from "here's where the funnel breaks" to "here's the redesign." The data pointed at where the problem lived, but treating that as a why would have meant designing against assumptions instead of evidence. Framing the deal results page as a decision-support experience, rather than a product catalogue reframed the entire brief: the goal wasn't to push customers toward any deal, but to help them understand their options and choose the right one with confidence, which is ultimately what protects completed switches rather than just clicks.

The proposed direction moved the page away from functioning like a product catalogue and toward a guided decision-support experience, a personalised recommendation paired with transparent comparison, rather than a flat list of packages customers had to decode themselves.

Comparison, without losing control; For customers who want to look beyond the recommendation, a compact comparison table on desktop (collapsing to expandable sections or stacked cards on mobile) preserves the ability to compare freely, so guidance never comes at the cost of control.

Personalised rationale; Where possible, the recommendation ties back to the customer's own quote inputs like household size, usage, stated needs, so it reads as reasoned rather than arbitrary.

Built-in reassurance; Installation expectations, what happens to the existing service, contract commitment, promotional vs. ongoing pricing, and any material fees are answered on the page itself, before the customer has to go looking for them.

I also restructured the information hierarchy around the customer's decision rather than the product catalogue's internal structure: recommendation -> deal name -> price -> speed -> who it's for -> why it's recommended -> key terms -> primary CTA.