“The website is outdated” can mean a lot of things. The team is tired of the font. The product has changed, but the page still sells its previous version. Sales reps keep explaining things the site should have explained for them.
A website redesign can make sense in any of those situations, provided everyone agrees on what it's supposed to accomplish. Changing the font is legitimate work. I just wouldn't promise it will stop customers from calling with questions about pricing plans.
What should people be able to do afterward?
I like starting with a fairly practical question: what will someone be able to do after the changes that they can't do now?
Choose a suitable option without help. Find the terms that apply to their company. Complete registration without calling an account manager. Understand what's being sold and how it differs from the next offer. Those answers give the project something to work toward.
“Look more current” can be a goal, too. Maybe a company is repositioning and no longer wants to look like the cheapest supplier. Then we need to understand what supports that new position: the offer, copy, photography, service. The visual layer can't be expected to carry all of that by itself.
If the only explanation is “we've been looking at this for five years,” I'd pause. A visitor may have opened the page for the first time today. They haven't been in a five-year relationship with your website.
When individual fixes stop adding up
The Soora Skin project involved positioning, packaging, visual identity, and the store's structure. It's a published case study, so I can name it here. The catalog, packaging, and brand needed a shared direction: people should recognize the same product across different touchpoints.
I was responsible for design; another team handled development. The case study contains no client sales metrics, and I won't add a sales increase just to make the story more persuasive.
What matters here is the scope. Fix the catalog grid alone, and the disconnect between the store and packaging remains. Change only the packaging, and the old store keeps telling its own story. A larger rebuild has a reason to exist when connected parts need to work together.
That doesn't mean every site needs the same treatment. If someone can't submit a form because of an error, that problem already has a location and a way to verify the fix. Rebuilding everything adds a lot of other work to it.
Less uncertainty means a clearer scope
Before redesigning a website, establish where you can observe a problem and where the team only has an explanation. That takes data, customer feedback, and a walkthrough of key user flows. I've covered the process in the article on website UX audits.
Then you can compare approaches, including the fairly ordinary costs of supporting, migrating, and explaining the changes.
| Approach | When it fits | What to account for |
|---|---|---|
| Fix one part | The cause is local and reproducible | Dependencies on the surrounding flow |
| Make incremental changes | The working foundation can stay | Supporting old and new patterns during the transition |
| Rebuild the experience | The new requirements no longer fit the existing structure | Content and data migration, cutover, and full validation |
Incremental work can develop an expensive tail, too. If every fix needs a workaround for the last one, the savings may exist only in the project name. “Let's never change the whole thing” doesn't appeal to me any more than “let's throw it all away.”
If the team already has a wish list, send me the site for a review. I'll look at which requests address the same problem and which have simply accumulated over time
Familiarity has value
An app redesign makes the cost of change particularly visible. Someone may perform a familiar task regularly, barely thinking about the controls. Move the action somewhere we consider more logical, and now they're searching for it.
That's no reason to preserve a poor decision forever. But habit is part of the product. Which tasks happen most often? Where will their sequence change? Can the transition happen gradually? How will people know their data and capabilities haven't disappeared?
Returning users don't owe an update their enthusiasm. They may just want to finish before lunch.
What I give AI and what I keep
AI can quickly produce something to evaluate: a different structure, draft copy, a component, a few states. I like being able to try a decision in the browser and discard it when it doesn't work.
Speed also makes it easy to believe an idea too soon. An untested assumption has already become a polished page, and people start discussing it as a nearly finished product. I try to keep the original question close: what were we trying to change in someone's behavior or understanding?
Payment logic, data access, and complex back-end work need specialists and separate validation. A quickly built interface doesn't answer those questions.
What needs to survive the migration?
I'd discuss website redesign cost alongside an inventory of what already works: pages, content, search URLs, analytics events, integrations, forms, permissions. The new version will need those things even if the presentation only shows the homepage.
A company website may include customer portals and workflows you wouldn't guess from its public pages. An estimate for new screens that excludes migration and validation leaves part of the project outside the quote. Someone will eventually pay for it in time or money.
I'm not offering a universal price range; it says too little without a defined scope. For stores, see e-commerce website redesign. For evaluating interface estimates, see app design cost.
Where I come in
I can connect the goal, structure, visual language, and implementation within the project's scope. First, though, I want to understand why the redesign is needed. Sometimes the answer leads to a substantial project. Sometimes it removes the need to redraw half the site.
Send me a link and tell me what isn't working for you. I'll review which changes are worth testing first and what supports that decision
Sources and limitations
The published Soora Skin case study was reviewed in its Russian-language version on November 4, 2026. The design scope is documented there; the original brief also identifies the separate development team. No sales outcome is claimed. The other recommendations describe my approach, without guaranteed improvements or fixed timelines.






