The honest headline numbers
SEO in the UK costs anywhere from £300 a month to £20,000+, which is a useless sentence on its own, so here's the shape underneath it. Prices track two things: who does the work, and how competitive your market is.
| Provider | Typical monthly | Hourly |
|---|---|---|
| Junior freelancer | £300–£800 | £30–£60 |
| Experienced freelancer | £800–£2,500 | £60–£150 |
| Small agency | £1,000–£3,000 | ~£60–£100 |
| Mid-sized agency | £3,000–£7,000 | £80–£150 |
| Large agency | £7,000–£20,000+ | £100–£300+ |
By scope: a single-location local business £300–£800, local with real content and technical work £500–£1,500, a growing SME or B2B lead-gen campaign £1,500–£5,000, ecommerce £2,000–£8,000+ (the median advertised ecommerce price in one UK dataset was £2,370), competitive national work £3,000–£10,000+.
Now the part the pricing guides don't lead with. A 2026 UK study that collected 129 actual advertised retainer prices found the median was £800 a month, while most agency-written guides insist meaningful work starts at £1,500+. Both things are sort of true: there's a large cheap market and a smaller expensive strategic one, and the average gets quoted by whoever it flatters. Treat every "from £X" as a floor and every tier table as an opening position, exactly like agency retainers in general.
What your money actually buys
SEO retainers are labour. So the only honest way to judge a fee is in hours and deliverables, and here's the arithmetic nobody itemises. One provider's own guide admits £1,000 a month typically funds 8 to 12 hours of genuine work. Across the benchmarks:
| Monthly fee | Real delivery hours | What fits |
|---|---|---|
| £500 | 3–8 | Google Business Profile upkeep, basic on-page fixes, light reporting |
| £1,000 | 8–15 | A focused local campaign, one content asset or a small technical backlog |
| £2,500 | 20–35 | Technical + content + selected outreach + actual strategy. The first band where a real programme fits |
| £5,000 | 35–60 | Multi-specialist: strategist, technical, content, digital PR, analytics |
Read that table backwards and the biggest SEO rip-off becomes obvious.
"A £500 proposal promising articles, links, PR and technical work is promising forty hours from an eight-hour budget. The gap is filled with templates, offshore production, or fiction."
The question that cuts through every proposal: "how many hours, by role, are assigned to us each month?" Honest providers answer it. And ask who actually touches the account, because the oldest trick in the industry is the senior consultant who does the sales call and is never seen again. Juniors doing production work is fine and normal; not knowing who does what is not.
How long before it pays?
Google's own guidance on hiring an SEO says four months to a year before benefit. The realistic shape by scope:
| Your situation | Early signs | Often-cited positive ROI |
|---|---|---|
| Local service business | 1–3 months | 5–9 months |
| Regional service business | 2–4 months | 6–12 months |
| Ecommerce | 3–6 months | 9–12 months |
| B2B with long sales cycles | 3–6 months | 12–20 months |
SEO genuinely needs runway, which makes a three-month minimum contract defensible. But the first 90 days produce output, not revenue, so judge month three on leading indicators, not rankings:
- Audit completed and priorities agreed
- Fixes actually implemented, not just recommended
- Tracking configured and owned by you
- Pages shipped or improved
- A work log you can inspect
"SEO takes time" is true; it is not a licence for three months of uninspectable activity. And do the worth-it maths on gross profit, not traffic: an accountant paying £2,500 a month (£30k a year) who gains 50 qualified enquiries and converts 10 at £5,000 first-year profit each made £50k against £30k, a 67% return before any retained-client value. If you can't sketch that calculation for your own numbers, start with the Pipeline Calculator, not a retainer.
The rip-off patterns
SEO can fail honestly: weak offer, brutal market, slow implementation. But these patterns aren't honest failure, and UK buyer guides, Google's own guidance and a decade of forum threads agree on them:
- Guaranteed rankings. Nobody controls Google. The "guarantee" is either zero-volume keywords, vanity terms, or small print.
- Bulk link packages. "500 backlinks for £299" buys risk, not authority. Demand every live URL and the right to veto placements.
- Report-only retainers. Ranking screenshots and "health scores" instead of a work log. A report is not delivery: ask what changed, who changed it, and what's next.
- Lock-ins without exits. Twelve-month minimums, buried auto-renewals, exit fees. Fair version: a three-month diagnostic, a documented plan, then a notice period.
- Agency-owned accounts. You must own Search Console, GA4, your Business Profile, the domain and every piece of content. Access is granted; ownership never leaves.
- AI slop sold as expert content. High article volume with zero access to your expertise is generic filler at premium prices. Ask where AI is used and who reviews it.
Doing it yourself (the part nobody selling SEO will tell you)
The cash cost of owner-led SEO is nearly nothing: Search Console, GA4 and Keyword Planner are free, Screaming Frog crawls small sites free, and a paid tool like Ubersuggest or Mangools runs £10–£25 a month if you want one. The real cost is your hours and the risk of working on the wrong things.
Here's my honest 2026 take, and it's first-hand: I do SEO for a living, and outside the day job I write and rank content for my own projects using AI as the production line. The loop I actually run, every time:
- Pull the keywords from real search data, and pick fights you can win (low competition, real intent)
- Read the pages that already rank, and find what they all miss
- Draft with AI against my own expertise and real experience, never instead of it
- Edit every line by hand, then publish and measure
The article you're reading was built with that exact loop. AI has genuinely moved the line: research, drafting and technical checklists that took a specialist a day now take a motivated owner an evening. What AI hasn't moved is knowing what to prioritise, real expertise in the content, and the patience to keep going for months.
So the honest split: if your market is competitive or your site has real technical problems, buy help. If you're a local firm with a decent site and more time than money, start yourself. You'll either succeed or become an unrippable-off buyer, and both outcomes pay. The vertical playbooks on this site (accountants, solicitors, IT firms and more) show exactly what to answer first.
Should you even buy SEO in 2026?
The AI-search complication
Fewer than a third of Google searches now end in a click, and pages cited by AI assistants increasingly aren't the same pages ranking top ten. Classic SEO still matters (it's the foundation AI answers are built on), but a rank-tracking-only programme measures less of reality every month. Ask any provider how they measure AI visibility and qualified demand, not just positions.
"Most of what's sold as GEO is good SEO wearing a new badge. Pay for the work, not the acronym."
Clear entities, structured data, genuinely useful answers, real expertise: that's the substance behind the AI-visibility pitch. A 10–30% retainer uplift for doing it properly can be legitimate; a proprietary "AI score" with no methodology is not.
The readiness test
SEO is the wrong first purchase for some businesses. It's not your channel yet if any of these is true:
- Nobody searches for what you sell
- You can't fund 6 to 12 months of it
- You need leads this month to make payroll
SEO is the compounding channel you build once something faster is keeping the lights on. The full hiring guide covers that decision properly, including when the right amount of marketing help is none at all.
Figures are 2026 UK benchmarks triangulated from independent sources: SaySEO's 129-price UK observational study (Aug 2026), the Ahrefs provider survey, SE Ranking's agency pricing and profitability surveys, Whito's UK cost index, SparkToro's zero-click research and Google's own SEO-hiring guidance, cross-checked against UK agency guides and practitioner discussion. Where a figure is provider-claimed or secondary, the text says so.