Why Your Website Redesign Keeps Not Fixing Anything

Most Series A-B teams that greenlight a website redesign ask one diagnostic question before they sign the statement of work: does this look right for our stage. It’s the wrong question. It’s answerable by any competent agency, which is exactly why the redesign gets built, ships, and then changes nothing about pipeline, win rate, or sales cycle length.

The right question is different: have we validated why a customer chooses us over the next-best alternative. That’s not a design question. It’s a positioning question, and it has to be answered before anyone opens Figma.

The Vendor Problem, Not the Execution Problem

When a redesign “doesn’t work,” the instinct is to blame execution: the wrong brief, the wrong timeline. But the failure usually isn’t in how the site was built. It’s in what the engagement was scoped to produce.

A brand or web agency’s deliverable is a site: information architecture, visual system, copy, launch. A positioning engagement’s deliverable is different in kind: a validated narrative of who you’re competing against in the buyer’s head, what you have that they don’t, and which segment of the market actually cares about that difference enough to self-select in. April Dunford’s 10-step positioning methodology, laid out in the 2019 edition of Obviously Awesome (a later edition consolidates the same work into 5 components), is the closest thing B2B SaaS has to a named standard for this work. It’s the framework this piece treats as a prerequisite to any customer-facing redesign, not a byproduct of one, even though Dunford’s own book frames positioning as foundational to marketing and sales enablement generally, not specifically sequenced before a website project. Competitive alternatives, unique attributes, value, target market characteristics, market category, and relevant trends all get worked out and written down before anyone talks about hero copy or a homepage layout.

Most redesign engagements never produce that artifact, and the reason is structural, not a matter of diligence. The vendor doing the design work isn’t the one who owns the account for validating positioning. Nobody on the project is tasked with going back to customers and asking why they bought, who else they considered, and what almost made them walk. So the positioning question gets answered informally, in a kickoff workshop, by whoever’s in the room, and it shows up as a paragraph in a creative brief instead of a signed-off artifact. The redesign then faithfully executes a positioning nobody actually tested.

What This Looks Like When It Ships Fine and Fails Anyway

Userlist is the case that shows what changes when the sequencing is inverted. Before touching their site’s visual design, the team applied Dunford’s 10-step method from Obviously Awesome and repositioned from “email automation for SaaS” to “customer messaging for SaaS.” That repositioning, not a redesign, is what let them name a competitor like Intercom directly on the page and have the value proposition land instantly. When they launched on BetaList, they got multiple unprompted comments praising how clearly the new positioning communicated what the product did. Nobody complimented the color palette. They complimented clarity, and clarity was a positioning output, delivered before the design work started.

Contrast that with what happens when a redesign is funded to fix a problem that isn’t a design problem. In one case ConvertCart describes on its blog, a jewelry brand spent $30,000 on a new brand identity and visual redesign, and conversion rates barely moved. The actual friction, by ConvertCart’s account, was checkout: over 60% of cart abandonment traced back to checkout friction, not to how the brand presented itself visually. Fixing that checkout flow, not the redesign, is what produced the 18% conversion lift ConvertCart credits to the fix. Treat the specific figures as illustrative rather than audited, since the case is unnamed and undated, but the pattern holds: the money went into the wrong layer of the funnel, and the actual fix was cheaper and more specific than a rebrand.

The same source describes a home decor retailer that redesigned its desktop site without prioritizing mobile, even though 80% of its traffic was mobile, and saw sales drop within a week of launch. Again, by the vendor’s own telling, nothing was wrong with the design craft. The redesign optimized an experience that wasn’t where the actual buyers, or the actual objection, lived.

The Fix Is a Gate, Not More Research

The instinct after reading cases like these is usually “we should do more research before the redesign.” That’s not quite it, and it’s the trap that keeps the same failure recurring under a different name. The fix is sequencing and ownership. Positioning has to exist as a distinct, prior deliverable, with its own owner and its own sign-off gate, separate from the design project and separate from whoever’s running it.

Concretely: someone on your team, or someone you’ve engaged specifically for this, needs to interview recently won and recently lost customers, map the competitive alternatives they actually considered, and write down the unique attributes that made the difference, before a single wireframe exists. That document gets approved on its own, by whoever owns revenue outcomes, not folded into a creative brief where it can be waved through in twenty minutes. Only after that artifact is signed off does the redesign brief get written, because now the brief has something true to say.

At Chalk Theory, this is why positioning work sits inside our core GTM and RevOps practice and is treated as a prior, separate deliverable, not a slide in the kickoff deck for Chalk Studio’s design engagements. If the redesign is the thing you’re funding, it’s worth checking what it’s actually being asked to fix first.