The short version: for a small site of five to fifteen pages, keyword research should take somewhere between half a day and two full days, roughly 30 to 90 minutes per page depending on how competitive the topic is. If it’s taking you three weeks, you’re not doing keyword research anymore, you’re avoiding writing.
Why “how long” is the wrong first question, but I’ll answer it anyway
Everyone wants a number. Fair enough, I’ll give you one. But the number only makes sense once you know what you’re trying to find out, which is covered in what keyword research really helps you answer. It’s not a treasure hunt for a magic phrase with 2,000 searches and no competition. It’s a way of confirming that the words you were going to use anyway are the same words your customers type into Google. That’s it. Once you frame it that way, the time it should take gets a lot shorter.
For a small site, here’s my working rule after doing this for over fifteen years and now rebuilding my own site from scratch in 2026:
- 1 to 5 pages: 2 to 4 hours total
- 5 to 15 pages: half a day to a full working day (4 to 7 hours)
- 15 to 30 pages: 1.5 to 2.5 days
- Anything bigger than that isn’t really a “small site” question anymore
Those numbers assume you already know your business and your customers. If you’re guessing at your own market as well as guessing at keywords, add half again.
The plumber who took three weeks and lost the launch date
I worked with a plumbing and heating client in Kent a couple of years back, a solid local business, six vans, decent reviews, needed a new six-page site: home, about, boiler repair, boiler installation, emergency callout, contact. The developer was ready to go. But the owner’s son, who was “handling the marketing,” decided the site couldn’t launch until the keyword research was “finished.”
He spent three weeks in Ubersuggest and Google Keyword Planner, building a spreadsheet with over 400 rows, comparing search volumes for “boiler repair near me” versus “emergency boiler repair Kent” versus “local boiler engineer” down to single-digit monthly search differences. The site sat unlaunched the whole time. In that window they missed the start of the autumn boiler-breakdown season, their busiest period, because the old site had already been taken down in preparation for the switch.
The finished spreadsheet was thorough. It was also almost identical to what I’d have written on a napkin in twenty minutes: five main service terms, one branded term, one local variant per page. The three weeks didn’t buy better keywords. It bought a worse launch date.
What the hours should go on
Here’s the breakdown I use with clients now, page by page:
- List every page and its job (15 to 20 minutes total). What is this page meant to do: sell a service, answer a question, capture a lead? That job comes before any keyword.
- Pull search volume for the obvious term (10 minutes per page). Google Keyword Planner, or Ahrefs or Semrush if you have access, gives you a rough monthly search number. Rough is fine. You don’t need it accurate to the visitor.
- Look at the actual top 10 results (15 to 20 minutes per page). Not the search volume, the results themselves. Are they big national brands you can’t outrank in year one? Are they thin, outdated pages you could beat with something better? This tells you more than any tool.
- Check three to five related phrases (10 minutes per page). Related searches at the bottom of the Google results page, “People also ask” boxes, and if you sell anything with intent close to advertising, the same logic used in keyword research for Google Ads campaigns works nearly as well for organic, just without the cost-per-click data.
- Write down the one main term and two supporting terms per page (5 minutes per page). Not fifteen. Two supporting terms. That’s the whole output.
Add that up for a six-page site and you get roughly 55 to 65 minutes per page, so five to six and a half hours total, spread across a day. Not three weeks.
The bit most keyword research guides won’t tell you
Here’s the part that gets skipped in most guides on this topic. For a small site, most of the traffic you get in year one will come from a handful of terms you’d have guessed correctly without any tool at all, plus a long tail of phrases you never researched because you can’t predict them in advance. I’ve watched this happen on my own site. I put real hours into the “planned” keyword list for a service page, and six weeks later Search Console showed me the actual queries bringing in clicks, half of which weren’t on my list, because real customers phrase things more oddly and more specifically than any keyword tool predicts. “AI consultant for a five person team” beat every polished variant I’d researched.
That means the honest use of your research time is closer to setting a direction than nailing a target. Spend the hours getting the general topic and intent right. Don’t spend them trying to guess the exact phrase, because Search Console will tell you the real answer for free in about six to eight weeks, and it will be more accurate than anything you paid a tool for.
When to spend more time than the guideline
The half-day to two-day range breaks down in a few situations, and it’s worth knowing which one you’re in before you start the clock:
- competitive niches. If you’re launching a small site in a crowded space like personal injury law or weight loss coaching, an extra half day spent reading the top-ranking pages is worth it, because the gap between “close enough” and “good enough to compete” is bigger.
- Multiple services with overlapping language. If your six pages could easily blur into each other (three different types of “coaching,” say), give yourself extra time to separate the search intent so you’re not writing five pages that all want to rank for the same phrase.
- You’re also running paid search. If Google Ads is part of the plan, do that research alongside the organic pass rather than twice, since a lot of the volume and intent data overlaps.
- AI search is changing your traffic mix. Worth reading whether keyword research is still important in the age of AI search before you assume the old rules still apply exactly as they did in 2022. The short version there is that intent still matters even when the search box has disappeared into a chat window, but the volume numbers matter a bit less than they used to.
None of those situations justify three weeks either. They justify an extra hour or two, not an extra month.
What happens if you outsource it instead
If you’d rather pay someone to do this and get your time back, that’s a fair trade, and the going rate for a small site keyword research job is covered in detail in how much to charge for keyword research services. As a rough guide, a freelancer doing a proper job on a ten-page site should be quoting you somewhere between two hundred and six hundred pounds for the research alone, not the writing. If someone quotes you two thousand pounds and three weeks for a six-page local business site, ask them what they’re spending the time on, because based on the plumber story above, the honest answer is often “building a spreadsheet nobody will read past page one.”
A simple timer test
If you want a practical way to keep yourself honest: set a kitchen timer for 45 minutes per page and stop when it goes off, even if you feel like you’re “just about to find something.” Write down whatever you have. In my experience, what you have after 45 minutes is almost always what you’d have had after four hours, just with less second-guessing attached to it. The extra hours mostly get spent on comparing near-identical numbers rather than making better decisions.
Frequently asked questions
How many hours should keyword research take for a five-page website?
Budget two to four hours total for a five-page site, roughly 30 to 50 minutes per page including a quick check of search volume, a look at the top results, and writing down one main term plus two supporting phrases per page.
Is it normal for keyword research to take weeks for a small site?
No. If keyword research for a site under 20 pages is stretching into weeks, the work has usually turned into indecision rather than research, comparing similar numbers repeatedly instead of moving to writing and publishing the pages.
Should I do keyword research before or after writing the page?
Do a light pass before writing, just enough to confirm your main term and intent, then write the page, then check Search Console after six to eight weeks. The real data from actual visitors is more useful than any amount of extra pre-research time.
Do I need a paid tool like Ahrefs or Semrush for a small site, or is the free Google Keyword Planner enough?
For most small sites under 20 pages, Google Keyword Planner plus a manual look at the top 10 search results is enough. Paid tools save time if you’re doing this often or for clients, but they don’t change the core decision for a handful of pages, so don’t delay launch waiting for a subscription.
