You are wondering whether to spend real money on a website when AI answers half the questions people used to click through for, and everyone you know runs their business off Instagram. It is a fair question and you should be suspicious of anyone in our line of work answering it with an immediate yes. So we will not. There are businesses that genuinely should not make a website their next spend, and we will name them early, because an argument that concedes nothing is not an argument. What is true is that something specific changed in the last two years, it is measurable, and it changes what a website is for rather than whether you need one. This is aimed squarely at you as the owner rather than at people wondering whether to learn the trade, which is a different question with a different answer.
Key Takeaways
Pew Research tracked 68,879 Google searches from 900 US adults in March 2025 and found that users who saw an AI summary clicked a traditional result on 8% of visits, against 15% for those who did not. Sessions ended on 26% of pages with a summary against 16% without.
In that same study, 18% of searches produced an AI summary at all, rising to 60% for question-format queries and falling to 8% for one and two word searches. So this hits informational questions hardest and name-and-service searches least.
BrightLocal's 2025 survey of 1,026 US consumers found only 42% now trust online reviews as much as a personal recommendation, down from 79% in 2020, while 74% check two or more sites before deciding.
Not the aesthetics. The domain, the content, the indexable pages and the ability to keep operating through somebody else's policy change. That is what a social profile cannot give you, and it is the whole argument.
Are Websites Still Relevant in 2026?
Yes, and less universally than five years ago, which is a more useful answer than a flat one.
What has changed is that a website is no longer the automatic destination of every search. It used to be that being findable and having a page were nearly the same thing. Now some questions get answered before anybody arrives anywhere, some businesses get chosen from a map listing without a click, and a real share of buying research happens inside apps. If your website's only job was to be the place a search sent people, that job has genuinely shrunk.
What has not changed is what happens once somebody is deciding between you and two competitors. That is still a moment where somebody wants to look at your work, read what you actually do, check that you cover their area, and see whether you look like a real business. A small account making the point put it plainly: social media might help someone find you, but your website helps them decide whether to contact you. That is not evidence of a trend and we are not offering it as one, but it is the right division of labour.
So the honest version is that a website's role narrowed and deepened. Fewer people arrive by accident. The ones who arrive matter more.
Where a Website Is Genuinely Not Your First Spend
Three situations, and in all three we would tell you to spend the money elsewhere.
You do not have a settled offer yet. If you are still working out what you sell, to whom, at what price, a website is a monument to a decision you have not made. You will rebuild it. Get the offer right first, on the phone, with real customers.
Your business runs entirely inside one marketplace or platform. If every job comes through a trade platform, a food delivery app, or a marketplace that owns the transaction, your effort is better spent on your standing inside that system. A website is a hedge against that platform, which is a real reason to build one, but it is not urgent while the platform is working.
You are booked solid on word of mouth. A trade with a six-week waiting list built on referrals does not have a demand problem, and a website will not fix a capacity problem. A single page confirming you exist, with your number on it, is the correct spend. Anything more is buying something you cannot use.
There is a fourth case worth naming, which is different in shape. If you have a website and it is not bringing in work, the answer is often not a better website. It is that nobody can find it, or that the pages do not say enough for anyone to decide. That is a findability problem or a content problem, and spending on a redesign to solve it is the most common expensive mistake in this whole category.
What Actually Changed: Fewer Clicks From the Same Searches
This is the measurable part, and it is worth getting exactly right rather than shouting about.
Pew Research Center tracked 68,879 unique Google searches from 900 US adults during March 2025, using actual browsing data rather than a survey of opinions. The findings, in their words: users who encountered an AI summary "clicked on a traditional search result link in 8% of all visits," while those who did not encounter one "clicked on a search result nearly twice as often (15% of visits)." Users ended their browsing session on 26% of pages carrying an AI summary, compared with 16% of pages with only traditional results. Links inside the AI summaries themselves were clicked on 1% of visits.
That is a real effect, honestly measured, and anybody telling you nothing has changed is not paying attention.
Now the part that gets left out when this study is quoted at you. In the same data, 18% of all searches generated an AI summary at all. The breakdown matters more than the headline: 60% of question-format searches produced one, 53% of searches of ten words or more, and only 8% of one and two word searches. Read that against how people actually look for a business. "Emergency plumber Fairfield" is a two-word-shaped query with local intent. "Why is my radiator making a knocking sound" is a question. The first is the one that turns into work, and it is the least affected.
| Finding | The figure | What it means for a business site |
|---|---|---|
| Clicks with an AI summary present | 8% of visits | Informational traffic is genuinely thinner than it was |
| Clicks with no AI summary | 15% of visits | The gap is roughly half, not the collapse it gets described as |
| Searches producing a summary | 18% of all searches | Most searches were unaffected in that window |
| Question-format searches producing one | 60% | Your explainer content is where the loss lands |
| One or two word searches producing one | 8% | Name-and-service searches, the ones that convert, least affected |
| Clicks on links inside the summary | 1% of visits | Being cited in a summary is visibility, rarely a visit |
The strategic read is not "websites are finished." It is that traffic from broad informational questions is worth less than it was, and traffic from people looking for a specific service in a specific place is worth what it always did. Which means the pages worth having are the ones about what you actually sell.
Why Falling Trust in Reviews Puts More Weight on Your Own Pages
Here is a change that runs the other way and gets almost no attention.
BrightLocal's 2025 local consumer review survey, covering 1,026 US adults, found that only 42% of consumers now trust online reviews as much as a personal recommendation, down from 79% in 2020. Meanwhile 84% still use Google to find reviews, and 74% check two or more sites before deciding.
Sit with that combination for a second. People are still looking. They are looking in more places. And they believe what they find substantially less than they used to. Something has to close that gap, and a five-star average by itself no longer does it.
What closes it is specific, checkable detail about your actual work, which is a thing you can only publish somewhere you control. Photographs of jobs you did. A page about the exact service they need, with what it involves and what it costs. Your service area stated plainly. The awkward questions answered before they have to ask. That is a website doing the one job nothing else can do for it, and it is more valuable in a low-trust environment than it was in a high-trust one.
The Durability You Are Actually Buying
If we had to reduce the case for a website to one sentence, it would not be about design. It is that the website is the part of your web presence that cannot be taken away, retitled, throttled or deleted by a company you have no relationship with.
Your domain is yours, renewed by you, pointing wherever you say. Your content is yours, in your accounts, exportable and reusable. Your pages are indexable, which means they can be found by something other than the app they live in. And your ability to trade does not depend on an algorithm decision, a policy update, a suspended account, or a platform changing how much of your own audience it will show you. Every business owner who has lost a social account, or watched their reach quietly fall by two thirds, understands this line item immediately. Everyone else finds it abstract until the day it isn't.
That is also why the same reasoning shows up in why we build sites you own outright rather than on rented systems. The argument is not that rented platforms are bad. It is that if the only version of your business that exists online is inside somebody else's product, you do not have a business asset, you have a tenancy.
What a Website Has to Do Now to Be Worth Paying For
The bar moved. A brochure with a phone number was defensible in 2015 and is not now, because the thin informational traffic that used to reach it has gone. Here is what earns the money in 2026.
A real page per thing you sell. Not a services list with one sentence each. A page somebody could decide from. This is the single highest-return change on most existing sites and it is content work rather than design work.
Content that answers rather than teases. Google's guidance on AI features in search is deliberately unexciting on how to be picked up: the same foundations apply, helpful content and crawlable pages, with no special markup to add. So being readable by answer engines is not a new discipline you have to buy, it is the old one done properly.
Pages a machine can read. Real links in the HTML, content present in the served page rather than assembled by scripts after arrival, facts stated in structured data. If a crawler cannot read your site, neither can the systems now answering questions on your behalf.
Speed and accessibility, measured rather than assumed. Core Web Vitals give the thresholds. A free checker will tell you about contrast and unlabelled fields. Both are cheap and both are commonly failed.
Something only you could have published. Your photos, your prices, your process, your area. In a world where generic explanation is free and instant, the only content with any scarcity value is the specific kind.
One line on the other question, since it dominates this search: whether web design is a good career in 2026 is a genuinely different topic, and most of what you will find for this phrase answers that rather than yours. The closest thing to a considered take among those results argued that web design is not dying, but that the easy middle layer is collapsing, with value shifting from building websites toward helping businesses make better decisions online. That happens to be the same conclusion from the buyer's side: paying for pages is worth less than it was, and paying for judgement about what the pages should say is worth more.
Is a website your next spend, or is something else?
1. Do you have a settled offer you have sold to real customers?
2. Where does most of your work come from right now?
3. Could you take on more work next month if it arrived?
4. Does your site have a page for each service you want more of?
5. If your main social account were suspended tomorrow, could customers still find and contact you?
6. When somebody is choosing between you and a competitor, what can they look at?
Pick an answer to begin.
If your answers clustered in the first column, a website is a straightforward yes and the question is only what it should say. If you answered "entirely referral" and "at capacity," you have permission to not do this yet. If the last two answers were in the bottom row, the case is stronger than you think and it is about durability rather than marketing.
Frequently Asked Questions About Whether Web Design Is Still Worth It
Are websites still relevant in 2026?
Yes, with a narrower role than before. Fewer people arrive from broad informational searches, because AI summaries answer many of those without a click, and Pew's tracking found clicks roughly halved when a summary was present. What has not changed is the deciding moment: when somebody is choosing between you and a competitor, they want to look at your actual work and read what you actually do, and that is something you can only publish somewhere you control.
Do I still need a website if I have social media and a Google Business Profile?
You need something you own, and those two are not it. A profile can be suspended, restricted, or shown to fewer people, and none of those decisions are yours. A profile is also a summary, not a place where somebody can read enough about a specific service to decide. Running both is the sensible answer: the profile helps people find you, your own pages do the convincing.
Will AI replace websites?
It is changing how people reach them rather than removing the need for them. Answer engines have to read something to answer with, and the businesses being described accurately in those answers are the ones whose own pages state their facts clearly. In Pew's data, links inside AI summaries were clicked on only 1% of visits, which means being cited is visibility rather than traffic, and visibility still needs a page behind it.
Is a website worth the money for a small local business?
It depends on where your work comes from and whether you have room for more. If customers search and compare, yes, and the return usually comes from pages about specific services rather than from the design. If you are booked out on referrals, a single page with your details is enough, and anyone selling you more than that is not looking at your situation.
Will AI replace web designers?
The tools have made producing pages much easier, which has genuinely compressed the value of assembling a site from parts. What has not been automated is deciding what a business should say, to whom, and in what order, and that was always the part that determined whether a site worked. It is worth noting that this question, and this search generally, is mostly being asked by people considering the career rather than by business owners.
Are web designers in demand in 2026, and should someone learn it?
That is a career question rather than a business one, and it is what most results for this phrase actually answer. The short version from practitioners in those results is that straightforward site building has become crowded and cheap, while work that involves judgement about a business's online decisions has not. We are not the right source for career advice and will not pretend to be.
Final Thoughts
Something real changed, and it is measurable rather than rhetorical: when an AI summary appears, clicks roughly halve, and sessions end more often. It also applied to 18% of searches in the study window, concentrated on question-shaped queries rather than the name-and-service searches that actually turn into work. At the same time, trust in online reviews has fallen from 79% to 42% in five years while people check more sources than ever, which puts the burden of convincing back onto pages you control. Put those together and a website in 2026 is worth less as a destination and more as the place where somebody decides.
The practical benefit of thinking about it this way is that it tells you what to spend on. Not a redesign. A real page for each thing you sell, content specific enough that only you could have written it, pages a machine can read, and a domain and content in your own name so that no platform decision can take your presence away. That list is cheaper than a rebuild and it is where the return actually is.
If you want an honest read on whether this is your next spend, tell us where your work comes from now and whether you could handle more of it. At Web Leveling we build hand-coded sites for businesses whose website has to bring in work, and we will tell you plainly when the answer is that you should fix your service pages, sort out your local visibility, or simply do nothing this year. Tell us what is happening and we will reply in plain English within one business day. Whatever we build ends up in accounts registered to you, because the durability is the point and it does not work if we are holding it.
Terms
The terms in the websites-versus-AI argument
Tap a term to see what it means.
AI summary. The generated answer at the top of some search results. Appeared on 18% of searches in Pew's March 2025 tracking, and on 60% of question-format ones.
Zero-click search. A search that ends without anybody visiting a website. Not new, but more common where summaries appear.
Indexable. Able to be read and stored by a search engine. Content inside an app usually is not, which is most of the difference between a profile and a page.
Answer engine. Any system that answers a question directly rather than returning links. It has to read something to answer, which is why your pages still matter.
Platform risk. The chance that a company you do not control changes your reach, your rules, or your access. The thing owning a domain and your content protects against.
Structured data. Facts about your business stated in a form a machine reads without guessing. How you avoid being described inaccurately by something summarising you.