You get web design clients by expanding the accounts you already have, asking for referrals you are currently not asking for, and prospecting the redesign market rather than the businesses with no website at all. The survey data is consistent on this: repeat work produces more of a designer's business than referrals do, and the offline segment most outreach targets is both small and largely unconvinced that it needs you.
Almost every guide on this subject optimizes for net-new acquisition. The data points somewhere less exciting and considerably more profitable. Below is what the surveys actually show about where design work originates, which market is worth prospecting, and which claims to cut from your pitch before a prospect checks one. If you also sell marketing services, the broader playbook on how to get SEO clients covers the adjacent channels.
AI-generated audio summary of this article, created with Google NotebookLM.
Repeat work outperforms referrals in the data
"Referrals are how everyone gets clients" is the most repeated claim in freelance business advice, and the numbers are softer than the folklore.
In a 2025 survey of 267 freelancers, only 33% got more than half their business from referrals, and 42% got less than a quarter from them — while 77% got more than half their work from repeat clients. One caveat worth stating: that sample skews toward writers and editors rather than designers, so read it as evidence about freelance services generally.
Even discounted, the direction is a useful correction. The engine is retention and account expansion, not a referral lottery. For web design specifically that means maintenance and hosting retainers, iterative improvement and conversion work after launch, and the second site — the subsidiary, the landing-page family, the rebuild of the section nobody scoped the first time.
There is a structural reason this gets neglected: expansion work is unglamorous and nobody writes listicles about it. It is also the cheapest revenue you will ever acquire.
The referral bottleneck is the ask, not the willingness
When referrals do fire, the constraint is not enthusiasm. Buyer research found 69% of clients were very willing to recommend their provider — and roughly three-quarters of the time, nobody had asked them.
That reframes the whole tactic. A "referral system" is not a campaign to make clients like you more; it is the habit of asking at the moment the work lands well.
The same research found something that widens the pool considerably: 81.5% of firms had received a referral from someone who was never a client, typically triggered by hearing the person speak, reading their writing, or interacting with them online. Referral generation is not limited to past clients — it extends to anyone visible to the people your buyers ask. For web design, that circle is other agencies, developers, copywriters, and SEO firms whose clients need sites they cannot build.
Stop prospecting the businesses without websites
The most misdirected prospecting effort in this industry comes from a stale statistic.
"27–28% of small businesses don't have a website" is still quoted constantly. It comes from surveys published in 2021 and 2022 — and the same corporate parent's 2025 research now reports that 83% of US small businesses have a website, up from 64% in 2018. The offline share is 17%, not 28%. Anyone still citing the old figure is citing a number their own source has revised by roughly ten points.
The remaining 17% is worse than small. More than a third of those businesses say they do not believe a website is relevant to what they do. You would be selling the category before selling yourself, to the segment least persuaded that the category exists.
The redesign market is the real one. Measured through Internet Archive redesign intervals, 50 top marketing brands averaged 2 years 1 month between redesigns, while 78 agency clients averaged 6 years 4 months. Those are very different populations, which is exactly the useful part: marketing-led organizations refresh often, typical small businesses rarely. Target the former for cadence and the latter for the moment their site visibly stops matching the business.
Do not compete with the builder tier
The same 2025 survey shows how small business sites get made: 41% with DIY builders, 34% on WordPress or Shopify, and only 12% custom-built. Separately, 45% outsourced the work to an outside partner while 37% built it in-house.
A designer whose offer is "I will build you a website" is competing with a subscription product that costs less per year than an hour of their time. That is not a positioning problem to solve with better copy; it is the wrong race.
The 45% who outsource are the market, and they outsource because they want an outcome the builder tier cannot produce — a site that converts, integrates with the systems they already run, loads fast enough to rank, or fixes a specific commercial problem they can name. Lead with the problem. The build is the mechanism, not the offer.
Make outreach multi-touch and specific
Cold outreach works for designers, but rarely in the form most people attempt it.
An analysis of 12 million outreach emails found 8.5% received any reply, one follow-up increased replies by 65.8%, personalizing the subject line lifted response 30.5%, personalizing the body lifted conversions 32.7%, and contacting several people at the same organization lifted response 93%. Two honest caveats: the study is from 2019, predating the current bulk-sender rules, and it covers link-building outreach to publishers rather than sales outreach to business owners. Treat the mechanics as transferable and the 8.5% headline as optimistic.
The transferable lesson is that single-touch generic outreach underperforms by a wide margin, and both fixes are free. A teardown of a specific page on the prospect's own site, followed up once, beats a beautifully written introduction sent alone.
Retire the three statistics in your own sales deck
This is the uncomfortable part, and it applies to work you may already have shipped.
Three claims appear in a large share of web design pitches, and none survives being traced:
- "75% of users judge a company's credibility on its website design," credited to Stanford. The Stanford web-credibility study it is attributed to contains no such finding. That study was a self-report questionnaire, was not peer reviewed, and is from 2002 — describing the pre-mobile, pre-social web. The related Fogg research found visual design was mentioned in 46.1% of coded comments, which is a different measurement that appears to have mutated into the 75% figure.
- "94% of first impressions are design-related." Cited variously to Stanford, Forbes, or "a British study." No primary publication matching the claim is locatable.
- "Businesses redesign every three years." An agency went looking for the source and published its failure to find one. The figure appears to be a corrupted retelling of the Internet Archive analysis referenced above, which measured something narrower on a small sample.
Cutting these costs you nothing, because the underlying arguments stand without them. It protects something valuable: a prospect who checks one claim and finds it hollow will discount everything else you said. Buyers increasingly ask an AI to verify what they read, and untraceable statistics are exactly what that process surfaces.
Turn the teardown into a proposal
The prospecting method that fits all of the above produces one specific observation about one specific site. The conversion step is turning that into a document the owner can approve while the observation is still fresh.
That step is where most design pipelines leak. Writing a tailored proposal for every teardown is slow, so designers send generic ones, and generic proposals compete on price against the builder tier. The fix is to make the tailored version cheap: Pitchko turns a prospect's own URL into a personalized proposal in minutes, so the teardown becomes the opening of the document rather than an email nobody answered. For what belongs inside it — scope, revision rounds, and who is responsible for content — see the web design proposal guide.
Expand the accounts you have, ask the clients who would happily recommend you, and prospect the businesses whose sites already exist. That is a duller strategy than most advice offers, and it matches what the data says actually works.
