Why redesigns get chosen so often

When revenue softens, the storefront is the first thing everyone looks at, because it's the only part of the eCommerce business the whole team can see. Nobody opens the fulfillment settings in a leadership meeting. Nobody scrolls the tax and shipping configuration. But everyone has an opinion about the homepage.

So the conversation converges on appearance — and appearance is rarely the constraint. A store can look dated and convert well. A store can look beautiful and lose money on every order it takes.

A redesign changes what the store looks like. Most performance problems are about what the store does.

Diagnose before you design

Before committing budget, get specific about which number moved. "Sales are down" is not a diagnosis. Sessions, conversion rate, average order value and repeat rate fail in different ways, for different reasons, and each points at a different fix.

  1. 01

    Sessions fell

    This is a traffic problem — paid efficiency, organic visibility, email cadence or channel mix. A redesign does not create demand.

  2. 02

    Sessions held, conversion fell

    Look at traffic quality first, then at the specific step where people leave. A conversion decline concentrated in checkout is not a homepage problem.

  3. 03

    Conversion held, AOV fell

    This is merchandising, pricing, bundling or promotional mix. Usually solvable in weeks, without touching the theme.

  4. 04

    New customers held, repeat fell

    Retention, lifecycle messaging, subscription health or product experience. The storefront is barely involved.

Five problems a redesign won't fix

  • Unclear positioning. If the brand can't say what it is in one sentence, a new grid won't say it either.
  • Merchandising decisions nobody has made. Which products lead, which bundle, which get promoted, which should quietly disappear.
  • Operational friction. Shipping cost surprises, inaccurate inventory, slow fulfillment and returns pain are felt after the design ends.
  • Traffic quality. Buying cheaper, worse-matched clicks depresses conversion no matter how good the page is.
  • No owner. If nobody is accountable for eCommerce performance week to week, the new site will drift exactly like the old one did.

When a redesign genuinely is the right call

Sometimes it is. Reasonable triggers include:

  • The brand has materially changed and the site describes a company that no longer exists.
  • The theme is unmaintainable — years of stacked edits, dead apps and orphaned code.
  • Core Web Vitals or mobile experience are structurally poor and can't be patched.
  • The catalog model changed: new categories, subscriptions, B2B, bundles or a second market.
  • Accessibility gaps are baked into the markup rather than sitting on the surface.

Notice that each of those is a structural reason, not an aesthetic one. When the structure is the problem, rebuilding is cheaper than continuing to patch.

A better sequence

  1. 01

    Name the constraint

    One metric, one step, one system. Write it down before anyone opens Figma.

  2. 02

    Fix what's cheap and reversible

    Navigation labels, PDP content, shipping messaging, promotion clarity, page speed. Weeks, not quarters.

  3. 03

    Measure honestly

    Decide in advance what would count as improvement, and over what window. Give the change enough time to be real.

  4. 04

    Then decide about design

    If the constraint persists after the cheap fixes, you now have a much sharper design brief — and a much better chance of the rebuild paying for itself.

The point isn't that redesigns are bad. It's that the decision to rebuild should be the conclusion of an argument, not the opening move.