Case · E-commerce
Hans Audio: not a new webshop.
The request was a new store, with a budget to match. Research put the problem in two steps of the ordering process, plus in how products were found. What got built was a fraction of the plan.
What was asked, and what was going on.
The request
The enquiry was clear: the webshop needed replacing. It felt dated, revenue per visitor was slipping, and an agency proposal was already on the table. The question to me was whether that proposal was reasonable.
That is a good question, but not the first one. Before you can judge a proposal, it has to be clear which problem it is meant to solve. So the work started with establishing what was actually costing money.
What the research showed
Traffic was stable. Interest in products was stable. What stood out: visitors who had added something to the basket dropped out at an unusually high rate at two specific moments — and more on mobile than on desktop.
- Shipping cost only became visible at the last step, sending visitors back to recalculate.
- The order form behaved badly on phones: fields jumping, error messages you could not see.
- Products were hard to compare against each other, which in audio is the buying conversation.
- Some product pages were effectively invisible to search engines because of how variants were set up.
What a rebuild would not have solved
None of those four things gets better automatically with a new design. They could reappear in a new store, because they come from configuration choices rather than from the design. A rebuild also carried a risk the proposal did not mention: losing the visibility the existing product pages had built up.
The decision we made.
Three options were on the table. This trade-off was walked through with the owner before anything was built.
“Rebuild as proposed”
Rejected
Highest cost, longest timeline, and the risk of lost visibility. Did not specifically address the four problems found.
“Platform migration”
Rejected
The platform was not the constraint. Configuration was. Migrating would carry the same choices across.
“Targeted repair plus product structure”
Chosen
Four concrete interventions, measurable step by step, with existing visibility preserved. Roughly a fifth of the original budget.
The decision was not “do not build”. It was: build only what solves a demonstrated problem.
What was done.
Revised checkout. Shipping cost visible from the first step.
What it returned.
The four interventions shipped separately, so the effect of each was visible on its own. That is the main difference from a rebuild: when you replace everything at once, you cannot tell afterwards which part made the difference.
The most important outcome is not a percentage. It is that the owner now knows which step of his sales process is sensitive, and therefore where the next improvement sits. The remaining budget went into product photography and content — work that contributes to sales directly.
In numbers
- Budget
- Around 20% of the original proposal
- Timeline
- Weeks rather than months
- Visibility
- Preserved; no migration loss
- Next
- Remaining budget to photography and content
- Figures
- Final results pending client approval
Is there a proposal on your table?
Send it over. I will tell you which problem it solves, and which it does not.