Redesign or CRO? How to Know Which One Your Business Actually Needs

This is the most common question we get, and it usually arrives phrased the same way: "Where am I going to get the most bang for my buck?"

Underneath that question is usually a second, less comfortable one: "Am I thinking too short term?" A redesign feels like the bold, long-term move. CRO feels incremental — safer, but maybe not enough. Or the reverse, depending on which vendor last presented to your team.

Both framings are wrong, because redesign and CRO aren't points on a spectrum from tactical to strategic. They're answers to different questions. The mistake isn't choosing wrong between them — it's not diagnosing which question you're actually asking.

Here's the framework we use with clients.

There's a third case this framework doesn't cover: a redesign proposed because a replatform is already underway. That's a different decision with its own tradeoffs — see should you redesign while you replatform?

Start With Why You're Asking

There are essentially three situations that bring merchants to this decision. Each has a different answer.

Situation 1: You're going through a rebrand

Answer: redesign. This one isn't close.

If the brand itself is changing — new positioning, new identity, new visual language, new voice — the site has to change with it. You cannot A/B test your way to a new brand. CRO operates by making incremental, testable changes to an existing experience; a rebrand requires a coherent, holistic expression of something new. Those are fundamentally incompatible modes of work.

Trying to deliver a rebrand through CRO produces the worst of both worlds: a site that's half old-brand and half new-brand for months, tests that get confounded by inconsistent presentation, and a brand launch that lands with no impact because it dribbled out one component at a time.

Redesign. Then run CRO against the new baseline once it's live.

Situation 2: Your site "feels outdated"

Answer: it depends — and the distinction here is the most valuable thing in this article.

"Our site feels dated" is the single most common reason merchants ask about a redesign, and it's the one where the diagnosis matters most, because "outdated" covers two very different problems with two very different solutions.

Ask: is it outdated in style, or outdated in standards?

Style and imagery — typography that reads as 2018, photography that doesn't match how the brand looks in its current campaigns, a color system that's drifted from the rest of your marketing, layout conventions that feel stale. This is aesthetic debt. It's real, and it affects brand perception and trust. It lends itself to a redesign, because you can't fix a stylistic system by testing individual components against it. Piecemeal aesthetic updates produce incoherence, which often looks worse than the dated-but-consistent version you started with.

Web standards and modern UX patterns — this is a different animal. Poor Core Web Vitals. Mobile experiences that technically work but weren't built mobile-first. Product pages missing the information architecture shoppers now expect. Filtering and search that don't behave the way people have been trained to expect by every other site they use. Accessibility gaps. Cart and checkout flows that lag current conventions. This lends itself to CRO, because these are discrete, measurable, testable problems with identifiable revenue impact. You can fix them one at a time, prove each fix, and compound the gains — without the cost, risk, and six-month timeline of a full redesign.

Here's why this distinction matters so much financially: merchants routinely spend redesign budgets to solve standards problems. They rebuild the whole site because the PDP wasn't converting, when a targeted program of CRO work would have fixed the PDP for a fraction of the cost — and left budget for the next five problems.

Be honest in the diagnosis. Pull up your site next to three competitors. If your reaction is "ours looks older" — that's style, and it points toward redesign. If your reaction is "ours is harder to use" or "ours is slower" — that's standards, and it points toward CRO.

Often it's some of both. In that case, sequence: fix the standards problems through CRO first, because they produce measurable revenue that helps fund the redesign, and because you'll learn things about user behavior that make the eventual redesign dramatically better informed.

Situation 3: You're trying to grow revenue

Answer: CRO. But the partner you choose will determine whether it works.

If the goal is growth — more revenue from the traffic you're already paying for — CRO is the right investment. This is the situation where the "am I thinking too short term?" anxiety is most misplaced, because a well-run CRO program is not a short-term play. It's a compounding one. Each validated win becomes permanent, and the learnings accumulate into a genuine understanding of how your specific customers behave.

A redesign, by contrast, is a single large bet placed largely on judgment. Sometimes it pays off enormously. Sometimes conversion drops and nobody can say why, because everything changed at once. If your primary objective is revenue growth, you generally don't want your outcome determined by one unvalidated bet.

But — and this is the part most CRO pitches skip — the partner matters more than the discipline.

There is a large category of CRO vendors whose entire model is testing velocity. They'll run a lot of tests. They'll show you a dashboard with a lot of activity. And a year later you'll have a marginally better button color, a handful of statistically noisy "wins," and no meaningful change to your business.

Testing volume is not a strategy. The value in CRO is in identifying the right lever to pull. That requires understanding your funnel as a system — where the actual friction is, which page is quietly costing you the most, how acquisition and retention interact, what shoppers in your specific vertical expect from a brand like yours. It requires forming a real hypothesis about the business, not generating test ideas from a listicle.

So when you evaluate a CRO partner, the question isn't "how many tests will you run?" It's "how will you decide what to test?" If the answer is a process for identifying the highest-leverage points in your funnel, you're talking to the right kind of partner. If the answer is a monthly test quota, you're buying activity, not outcomes.

The Cases Where the Framework Bends

A few real-world complications worth naming:

Your site is technically broken. If the foundation is genuinely failing — unusable on mobile, severe performance problems, an architecture that can't support what marketing needs — no amount of CRO will save it. Rebuild. You can't optimize a broken foundation.

You're replatforming. Don't stack a redesign onto a migration. We've written about this at length: bundling them multiplies risk and destroys attribution. Get to parity on the new platform, stabilize, then decide about design from a clean baseline.

Your traffic is too low to test. CRO requires statistical power. If you're not getting enough traffic to reach significance in a reasonable window, a testing program will frustrate everyone. In that case, apply CRO principles — best-practice UX, performance, standards compliance — through direct implementation rather than through testing, and revisit formal experimentation as traffic grows.

Leadership has already decided. Sometimes there's a redesign mandate from the top and the real question is how to make it succeed rather than whether to do it. Fair enough — that's a legitimate business reality. In that case, insist on two things: instrument thoroughly before launch so you have a clean baseline, and bring behavioral data into the design process so it isn't pure judgment. A redesign informed by CRO learnings is a far better bet than one informed by taste alone.

The Short Version

Rebranding? Redesign.

Site feels dated? Diagnose whether it's style (→ redesign) or standards (→ CRO). Most merchants assume style and are actually looking at standards.

Chasing growth? CRO — with a partner who leads with funnel diagnosis, not test volume.

Foundation broken? Rebuild first.

The most expensive version of this decision isn't picking the wrong one. It's not diagnosing at all — spending a redesign budget on a standards problem, or running a two-year testing program on a site that needed a rebrand.

At Anatta, we do both — and we've told plenty of clients that the thing they came asking for isn't the thing they need. Our funnel engineering work delivers an average 20–50% boost in conversion rate and a 35% increase in repeat purchases, and our redesign work has produced results like True Botanicals' move from a 4.1% to a 6.2% conversion rate. Our clients stay with us an average of three years, which tends to happen when the first recommendation is the honest one.

If you're weighing this decision, chat with our team — we'll help you diagnose it before anyone talks about scope.

Frequently Asked Questions

Is CRO cheaper than a redesign?

Usually, on a per-engagement basis — but that's the wrong comparison. The better question is cost per unit of measurable outcome. CRO produces attributable, compounding gains; a redesign is a larger single bet with a broader range of possible results.

How long before CRO shows results?

Meaningful results typically emerge over months, not weeks, because tests need statistical significance and the highest-leverage opportunities take time to identify properly. Any partner promising immediate lift is describing luck, not process.

Can I do a redesign and CRO at the same time?

Not effectively as parallel workstreams — a redesign changes the baseline your tests are measured against. The productive sequence is to use CRO learnings to inform the redesign, then resume testing against the new baseline once it's stable.

How do I know if my site is outdated in style or in standards?

Compare it to three competitors. "Ours looks older" is a style problem, pointing toward redesign. "Ours is harder to use" or "ours is slower" is a standards problem, pointing toward CRO. Check Core Web Vitals and mobile UX specifically — those are objective and frequently the real culprit.

What should I look for in a CRO partner?

How they decide what to test. A partner focused on identifying the highest-leverage points in your funnel will outperform one selling testing volume, every time.