Section 1
Why periodic full redesigns are usually the wrong model
The "redesign every few years" model has real problems. It's expensive and disruptive, a full redesign is a large project that consumes resources and attention. It's risky, replacing a whole site at once means you can break what was working (a redesign can reduce conversion if the new version is worse, and you've changed everything at once so you don't even know what hurt). It's often driven by the wrong reasons, boredom ("we're tired of how it looks"), fashion ("it looks dated"), or vanity rather than evidence that the current site is underperforming. And it throws away learning, the accumulated knowledge embedded in a site that's been optimized over time gets discarded for a fresh start. Meanwhile, the actual goal, a site that performs well and stays current, is usually better served by continuous improvement: evolving the site based on evidence (testing, data, real problems), which compounds learning, reduces risk (small changes, measured), and keeps the site current without the cost and gamble of periodic reinvention. So the contrarian answer: don't redesign on a schedule; improve continuously, and redesign only for a genuine reason. The principle: periodic full redesigns are expensive, risky, often boredom-driven, and discard learning, so continuous evidence-based improvement is usually the better model, with full redesigns reserved for genuine need. (This applies the conversion-optimization and web-strategy research cited across this library.) "Redesign every two to three years" sounds like discipline. Usually it's the opposite: an expensive, risky gamble driven by boredom, throwing away everything you learned about what works, and changing so much at once you can't tell what helped or hurt. The site that quietly improves every month, tested, evidence-based, compounding, beats the one reinvented from scratch every few years because someone got tired of the colors. Freshness isn't a redesign schedule. It's continuous improvement.
Section 2
The contrarian model (and when a redesign IS warranted)
The default: continuous improvement. Evolve based on evidence, improve your site continuously using testing, data, and real identified problems, so changes are justified and learning compounds. Small, measured changes, lower risk than wholesale replacement; you can tell what worked and keep what's working. Stay current incrementally, refresh and modernize gradually, keeping the site current without a disruptive reinvention. When a full redesign IS warranted (genuine reasons): Fundamental problems, the site has structural issues (badly broken, fundamentally not converting, technically obsolete) that incremental improvement can't fix. Major business change, your positioning, offering, or audience has fundamentally changed, and the site no longer fits. Genuinely obsolete, the site is truly outdated in ways that materially hurt (not just "looks a bit dated", actually failing on mobile, performance, or basic standards). Evidence, not boredom, there's real evidence the current site is underperforming in ways a redesign would fix. So the answer to "how often should you redesign?" is: not on a schedule, improve continuously, and redesign only when a genuine reason (fundamental problems, major business change, real obsolescence, evidence of underperformance) justifies the cost and risk. For most service businesses, that means rarely, and continuous improvement in between. (This position is my judgment built on the cited research.)
Section 3
Redesign vs. continuous improvement, in one view
The takeaway: the industry advice to redesign every two to three years is usually wrong, periodic full redesigns are expensive, risky (they can tank a site that was working), often driven by boredom or fashion rather than evidence, and they throw away accumulated learning. The contrarian, better model for most service businesses is continuous improvement: evolving the site based on evidence (testing, data, real problems) with small measured changes that compound learning and keep the site current without the cost and gamble of reinvention. Reserve full redesigns for genuine reasons, fundamental problems incremental change can't fix, major business changes, real obsolescence, or evidence of underperformance, not the calendar. So how often should you redesign? Not on a schedule: improve continuously, and redesign only when there's a real reason. For most service businesses, that means rarely. (The contrarian position is my judgment built on the cited research.)
Section 4
Execute This With AI
Step 1, Inputs. Note your site's current performance, what's prompting redesign thoughts (boredom? evidence?), and any fundamental problems. Step 2, Run the prompt: You are giving me a CONTRARIAN take on website redesigns. Periodic full redesigns (every 2-3 years) are usually wrong, expensive, risky, boredom/fashion-driven, and they throw away learning. The better default is CONTINUOUS IMPROVEMENT: evolve based on evidence (testing, data, real problems), small measured changes that compound learning and keep the site current. Reserve full redesigns for GENUINE reasons: fundamental problems incremental change can't fix, major business change, real obsolescence, evidence of underperformance, not the calendar. My site's performance: [DESCRIBE]. What's prompting redesign thoughts: [boredom/dated-look/ evidence/problems]. Any fundamental problems: [DESCRIBE]. Do four things: 1. Tell me honestly whether I need a full redesign or continuous improvement. 2. If it's boredom/fashion, talk me out of a risky full redesign and toward improvement. 3. If there ARE genuine reasons, confirm them and scope what actually needs redesigning. 4. Give me a continuous-improvement approach to keep my site current without reinvention. Don't redesign on a schedule; improve continuously, redesign only for a real reason. Step 3, The reason test. "Is a real, evidence-based problem driving this redesign, or is it boredom and 'it looks dated'? If the latter, improve continuously instead of gambling on a full redesign." Tools and expected output. Any frontier chat model, plus your performance data. Expect an honest redesign-vs-improve call, a boredom-check, genuine-reason confirmation and scoping if warranted, and a continuous-improvement approach. The QA discipline: distinguish genuine reasons (fundamental problems, major business change, real obsolescence, evidence of underperformance) from boredom and fashion, and default to continuous, evidence-based improvement, which is lower-risk and compounds learning. A full redesign should clear a high bar because it risks the site that's working. The model gives the contrarian counsel; honest evidence (not boredom) decides if a redesign is warranted. The industry advice to redesign every two to three years is usually wrong, periodic full redesigns are expensive, risky, often driven by boredom rather than evidence, and they throw away accumulated learning. The better model for most service businesses is continuous improvement: evolving the site based on evidence with small measured changes that compound learning and keep it current without the cost and gamble of reinvention. Reserve full redesigns for genuine reasons, fundamental problems, major business change, real obsolescence, evidence of underperformance, not the calendar. So how often should you redesign? Not on a schedule: improve continuously, and redesign only when there's a real reason. For most service businesses, that means rarely, and continuous improvement in between.
Section 5
Keep reading
Keep reading in the Conversion-Centered Design cluster and across the library: [The Anatomy of a High-Converting Service Page in 2026](/blog/the-anatomy-of-a-high-converting-service-page-in-2026), [Designing the "Why Us" Section That Actually Differentiates](/blog/designing-the-why-us-section-that-actually-differentiates), [The Dark Patterns Hurting Your Long-Term Conversion (Even Though They Work Short-Term)](/blog/the-dark-patterns-hurting-your-long-term-conversion-even-though-they-work-short-term). Also relevant: [Speed-to-Lead: Why Page Speed and Reply Speed Are the Same Conversion Lever](/blog/speed-to-lead-why-page-speed-and-reply-speed-are-the-same-conversion-lever), [The Service-Business Website Priority Stack: What to Fix First When Everything Needs Work](/blog/the-service-business-website-priority-stack-what-to-fix-first-when-everything-needs-work), [Website Redesign vs. Conversion Optimization: Which Do You Actually Need?](/blog/website-redesign-vs-conversion-optimization-which-do-you-actually-need).