Key takeaways
- Seed keywords are the short words or phrases you start keyword research from. You rarely target them directly; they produce the longer searches you do target.
- The best seeds come from your own business: what you sell, the words customers use, and the queries Search Console already shows for your site.
- A seed that is too broad drowns you in searches you cannot win or do not want. A seed that is too narrow returns almost nothing. Aim for the words a buyer would actually type.
- Synonyms need their own seed. A list built from one phrase only contains searches with that phrase in them.
- Test three to five seeds side by side before you plan pages, and keep the one or two that return the most relevant, reachable keywords.
Seed keywords are where every keyword list starts: a word or a short phrase that describes what you offer, which a tool or your own digging then grows into dozens of longer, more specific searches. Pick the wrong seed and everything after it is built on the wrong ground. Pick a good one and you can turn seed keywords into a keyword list with real searches you can plan pages around.
I have worked in SEO and digital marketing for more than 16 years, and when keyword research goes wrong for a small business, the problem is often at the very first step. The owner types one big generic word, gets a list they could never rank for, and gives up. This article explains what a seed keyword is, where to find good ones, how broad they should be, and how to test a few before you commit to a content plan.
What seed keywords are (and what they are not)
A seed keyword is the starting phrase of keyword research. It is usually one to three words long and names a product, a service or a problem: “dog grooming”, “boiler repair”, “wedding cake”. On its own it is often too broad or too competitive to target with a single page. Its job is to produce ideas.
From that seed you get the longer searches people really type: “dog grooming for anxious dogs”, “mobile dog grooming near me”, “how often should a dog see a groomer”. Those are the keywords you write pages for. Our guide to the long tail keyword generator method shows how those longer searches are built and filtered.
It helps to keep three terms apart:
| Term | What it is | Example (made up) |
|---|---|---|
| Seed keyword | The short phrase you start research from | dog grooming |
| Keyword idea | Any search the research turns up | dog grooming prices for large breeds |
| Target keyword | A keyword you choose for one specific page | puppy first grooming appointment |
How to find seed keywords: five places to look
Good seed keywords are rarely clever: they are the plain words between what you sell and what people type. These are the five places I look first.
1. Your products and services
Write down every product category and service you offer, using the plainest name possible. Not “bespoke canine aesthetics”, but “dog grooming”. If you would not say it out loud to a customer, it is probably not a good seed.
2. The words your customers use
Read enquiry emails, call notes, reviews and chat messages. Customers often describe the problem, not the service: “my dog’s coat is matted”, not “de-matting treatment”. Google’s SEO starter guide makes the same point: “Think about the words that a user might search for to find a piece of your content.” Its example: “some users might search for “charcuterie”, while others might search for “cheese board”.”
3. Your Search Console queries
If your site already gets some impressions, Search Console is one of the most reliable sources of seeds you have, because it is your own data. Google’s help page on the Performance report explains that the queries dimension groups your data by “the search query users typed”. Sort by impressions and note the short phrases that keep appearing inside longer queries. Those are your seeds.
Open the websites of three or four businesses like yours and read their navigation menus, category pages and service lists. You are not copying their content; you are noticing the names they give things. If every competitor has a page called “puppy grooming” and you do not, that is a seed worth testing.
5. Google autocomplete and People also ask
Type your candidate seed into Google and read the predictions before you press enter. Google’s page on how autocomplete predictions work says they “reflect real searches that have been done on Google”. When Google shows a People also ask box on the results page, it adds related questions around that topic. Both are free, quick checks that a seed matches the way people really search.
Choosing the right size: too broad, too narrow, about right
The most common mistake with seed keywords is size. A seed that is too broad produces a list dominated by searches you cannot win, or that have nothing to do with your business. A seed that is too narrow returns a handful of rows, or none at all.
Here is how that looks for a made-up dog grooming salon:
| Seed size | Example seed | What tends to happen |
|---|---|---|
| Too broad | dog | Searches about breeds, food, training and adoption. Mostly irrelevant, often dominated by big sites. |
| About right | dog grooming | Searches about prices, styles, at-home tips, mobile services and appointments. Mostly relevant. |
| Too narrow | hypoallergenic dog shampoo for poodle puppies | Very few or no searches contain the whole phrase. |
A quick test: could a customer who types this seed reasonably end up buying from you? “Dog” fails. “Dog grooming” passes. The long phrase might pass, but it is already a long tail keyword, not a seed.
Two more rules of thumb I use:
- Keep brand names out. A seed with your own brand name only finds searches for your brand. A seed with a competitor’s brand name finds their customers, which is a different project.
- Use the word forms customers use. “Dog groomer” and “dog grooming” produce different lists, so if you are unsure, test both.
A good seed keyword is the phrase your customer would say first, not the phrase your marketing team would like them to say.
How a seed keyword turns into a keyword list in SeedLens
Our free keyword research tool SeedLens is built around one, so it shows clearly why the choice of seed matters.
You enter a seed keyword, a word or a short phrase, and pick one of four countries: the United States, the United Kingdom, Canada or Australia. The PDF report then contains the seed keyword and up to 99 searches that contain it, with words added before, after or inside it. Each row carries Google search volume, cost per click, search intent and keyword difficulty from 0 to 100 with a plain label, plus a 12-month trend. The ten best opportunities, ranked by onwardSEO’s own Opportunity score, open the report, and your seed’s own row is marked.
That design has three consequences for choosing seeds:
- Synonyms need their own report. Because the list only holds searches that contain your seed, a report for “dog grooming” will not include “dog groomer near me” or “pet salon”. If customers use two different words, test both.
- The number of rows tells you about the seed. A report can hold up to 100 keywords; when fewer exist, it says how many were found. A short list is a sign the seed is narrow. A full list dominated by Hard and Very hard keywords is a sign it may be too broad for a new site.
- One seed, one country per report. You get one free report a month, so it is worth choosing the seed carefully, using the five sources above, before you run it.
Its limits: it is not a rank tracker (it shows what people search for, not where you rank), and the report is a PDF only, with no spreadsheet export.
Testing several seeds: a worked example
Here is a made-up example of how I would test seeds for a small business. Imagine a dog grooming salon in a UK town planning its next pages. Nothing below is real data.
Step 1: list candidates. From the services list, customer emails and a look at three competitors’ menus, the owner writes down five candidates: “dog grooming”, “dog groomer”, “puppy grooming”, “dog nail clipping” and “mobile dog grooming”.
Step 2: check each one quickly in Google. In this example, “mobile dog grooming” brings mostly predictions with town names, which suggests a local page rather than a seed for a whole list.
Step 3: run the strongest seeds, one per report. In this example the owner starts with “dog grooming” for the United Kingdom. Imagine the report comes back with the full 100 rows: many informational searches (“how to groom a dog at home”), a good number of commercial and transactional ones, and a difficulty mix with plenty of Easy and Medium rows. The next month they run “puppy grooming” and, in this made-up case, get a shorter list where most rows are Easy. Both are useful, for different pages.
Step 4: group and choose. The owner groups the rows by intent, as the SeedLens page suggests: an informational keyword needs a guide, a commercial one a comparison, a transactional one a product or service page. They end up with a puppy grooming service page, a guide to matted coats and a prices page, each built on keywords from their reports.
Comparing seed keywords side by side
- Is most of the list relevant to what you sell? If not, the seed is too broad.
- Does the list contain enough rows to plan more than one page? If not, the seed is too narrow.
- Is there a healthy share of Easy and Medium keywords for a site of your size?
- Does the intent mix match your goal (guides, comparisons or service pages)?
- Did a synonym you use with customers produce a different list worth a second report?
Common mistakes when choosing seed keywords
- Starting with industry jargon. If only your trade uses the word, the list will be short and full of other businesses.
- Treating the seed as the target. “Dog grooming” is a seed. The page you write is more likely to target something specific inside the list.
- Running one seed and stopping. One report shows one slice of how people search; two or three show the shape of your market.
- Ignoring intent. A seed can produce a list that is almost all informational when you need sales pages. Our guide to aligning content with search intent explains how to read intent before you write.
- Stuffing every variation onto one page. Google’s SEO starter guide says its “language matching systems are sophisticated” and warns that “keyword stuffing is against Google’s spam policies.” Group close variations onto one page and write naturally.
Once you have chosen target keywords from your seeds, the next useful step is knowing where you rank for them today. Our guide on how to choose keywords to track explains how to pick the ones worth checking.
FAQ
What is a seed keyword?
A seed keyword is the short word or phrase you start keyword research from, such as “dog grooming”. Research grows it into longer, more specific searches, and those are what your pages usually target.
How do I find seed keywords for my business?
Start with your own products and services in plain words, then add the words customers use in emails and reviews. If your site already has impressions, the queries in Search Console’s Performance report are an excellent source. Competitors’ menus and Google autocomplete fill the gaps.
How many seed keywords should I use?
For a small business, three to five candidates is a good start. Check them quickly in Google, run the strongest ones as separate reports, and keep the one or two that produce the most relevant, reachable keywords.
Is a seed keyword the same as a focus keyword?
No. A focus keyword is the main search a single page targets. A seed keyword is the starting point of research, and many focus keywords can come out of one seed.
Can a keyword idea generator replace choosing seeds myself?
A keyword idea generator saves time once you have a seed, but it cannot know your business. A poor seed gives a poor list however good the tool is.