Cannabis Keyword Research: How to Find What Buyers Search

Why cannabis keyword research works differently

Cannabis keyword research uses the same method as any other niche, but three things about this industry bend the usual tools. The first is the ad ban. Google's dangerous products or services policy keeps THC products out of paid search, and Google's Keyword Planner is built for advertisers. With so little ad activity around cannabis terms, its data for this category is thin and makes a poor primary source.

The second is volume. Ahrefs and Semrush estimate search volume from clickstream samples and ad data, and cannabis gets weak coverage from both. A question like "can you take edibles on a cruise" can show zero in a tool while your Search Console shows it landing on your site every week. Treat a blank volume cell as unknown and look for other evidence.

The third is who owns page one. Leafly, Weedmaps, AllBud and Healthline hold most bare strain names and the big comparison terms. A small dispensary or brand site will not outrank them on "blue dream" this year. What they cover badly is everything specific: strain plus format plus city, troubleshooting questions, local rules, the things your budtenders hear daily. With paid search mostly closed, that specific organic demand is one of the few open doors for meeting new customers at scale.

Where to find seed keywords

Seed keywords for a cannabis site come from your counter first and from software second. Budtenders hear the real questions all day: what helps me wind down without feeling foggy, what's the difference between a cart and a dab, do you have anything like Runtz. Ask staff to jot down every question for two weeks. That list is primary research no competitor can copy, and your contact form and review replies hold more of the same.

Autocomplete. Google's suggestions come from searches people have already run. Work in an incognito window, since suggestions are personalized, and remember Google suppresses some cannabis completions, so an empty dropdown does not prove zero demand. Run the alphabet method on each seed ("edibles a" through "edibles z", then the letter in front), then add question starters (how, why, can, is, does) and modifiers after the seed (for, without, near, vs).

People Also Ask and related searches. Open the People Also Ask box on a seed like "edibles" and keep clicking. Two levels deep typically yields 20 to 30 real questions. Record them as you go, because the box reshuffles between sessions. Related searches at the bottom of the page give you a second trail to follow a few hops out.

Reddit. Cannabis subreddits show the questions people were unsure how to phrase for a search box: dosing worries, travel questions, product confusion. A question that recurs monthly in threads is being typed into Google daily, often in the thread title's exact words.

A paid tool, last. Feed the survivors into Ahrefs or Semrush for related terms and ballpark volume.

Here's how one seed can play out. Run "live resin" through the process. Straight suggestions give "live resin vs distillate" and "live resin carts". The alphabet pass adds "live resin badder". Question starters produce "how to store live resin". Forty minutes might produce 47 phrases, 31 after deduping and 12 after dropping anything irrelevant to a licensed retailer: enough for two product-page rewrites and four blog posts.

Sorting keywords by buyer intent

Buyer intent decides the page type for every cannabis keyword, so tag it before anything else. Three buckets cover nearly everything, and the words inside the query usually give the bucket away.

• Learn. "What is live resin", "how long do edibles take to kick in". What, how and is mark these. They get explainers.
• Compare. "Live resin vs distillate", "best gummies for beginners". Vs, best and "for" plus a use case mark these. They get honest comparison guides.
• Buy. "Live resin carts denver", "edibles delivery near me", "runtz price". Buy, near me, delivery, price, deals, open now, a city or a brand mark these. They get category, product and location pages.

Commercial terms follow patterns, so generate them. Cross every category and strain you carry with buy, near me, delivery, price, deals and open now, then check which combinations show volume in a tool. Twelve products and a handful of modifiers produce ninety combinations in five minutes.

When the words leave intent ambiguous, search the term and let the results decide. A map pack plus Weedmaps and Leafly listings means a local buying query. Other retailers' category pages mean transactional. Listicles and blog posts mean the term is less commercial than it looks. "Best edibles" is the classic trap: buying words on an informational results page, so it belongs to a guide that links to the category.

It also helps to hear the voice behind a query. A connoisseur types terpene profiles and THC percentages, a first-timer types "what does an edible feel like", a deal hunter types "cheapest ounces". Add a persona column to the sheet and gaps in your coverage show up fast. Whatever the bucket, never point a decision query at a blog post. Someone searching "buy edibles online" who lands on a 900-word article hits the back button.

Strain, product and edibles keywords

Strain and product keywords are where cannabis demand concentrates, and each menu category speaks its own vocabulary. Give every product line its own research pass and tab, inside one master sheet that records which page owns each term.

Strains. Strain names are ready-made keywords. The review count on a strain's Leafly or AllBud page is a free demand proxy: thousands of reviews means steady searches, nine means a niche cross. Start with what you stock, because a strain page for something on your shelf ends in a purchase. Check aliases first. Gorilla Glue #4 became GG4 after a trademark dispute with the adhesive maker, and searchers still type both. Stack the strain with a format and a place ("wedding cake pre rolls", "runtz carts", strain plus your city), and add what only you have: your batch's tested THC from the COA, your grower's notes, what regulars say about it.

Edibles. Edibles research runs on three modifiers. Format comes first, and gummies pull most of the demand, so they earn their own category while mints and drinks start as sections. Dose is the spine, from 2.5mg microdose gummies through 10mg standards to 100mg bars, and every rung is its own keyword. Put the milligrams in the title tag: "10mg Watermelon THC Gummies" ranks for its exact demand where "Watermelon Gummies" ranks for nothing. Flavor and diet terms like vegan or sugar-free are small and nearly uncontested, and the onset questions form the category's biggest informational pool.

Concentrates, vapes and accessories. Concentrate shoppers search in dialect. Map demand on extraction method (live resin, rosin, distillate) and texture (shatter, badder, diamonds), and keep solvent and solventless on separate pages. Vape shoppers search carts, disposables and 510 batteries. Accessories ship nationwide and their buyers search by spec ("4 piece titanium grinder with kief catcher", "14mm male quartz banger"), so mirror those specs exactly in titles. CBD lines sort on spectrum, strength and format, which our guide to building a keyword list for a CBD product line covers in detail.

New products. A new SKU has no search history, but its strain and format do. Title a GG4 live resin cart the way people search, and publish three to six weeks before launch so indexing and the promotion peak together.

Effect language is where product research turns into a compliance problem. "CBD for sleep" and "edibles for pain" carry heavy demand, and the FDA has sent warning letters over exactly that claim language. Describe dose, ingredients, aroma and onset, and never promise an outcome on a product page. Our guide on how to avoid health claims in cannabis content shows where the line sits.

Local and near-me keywords

Local keywords for a dispensary are won partly on the site and partly off it. "Dispensary near me" is the clearest case: Google takes the searcher's location from the device and ranks nearby businesses on relevance, distance and prominence, the three local ranking factors it names. Writing "near me" into a page changes nothing. Your Google Business Profile, exact hours, a real location page and steady reviews win those searches, and spoken queries resolve the same way.

The research side is a grid. Write your terms down one side ("dispensary", "weed delivery", "dispensary open now", "medical dispensary") and your place names across the top, then fill the cells. City terms are crowded. Neighborhoods, districts, landmarks and storeless suburbs are where smaller shops win, using the names locals actually say.

Promote a neighborhood to its own page only when Search Console shows impressions for it and you can write a couple hundred words true only of that area: directions, parking, delivery windows. Pages that swap a city name into a template are what Google's spam policies call doorway pages. Delivery zone pages follow the same rule, each with its own minimums, windows and boundary streets, inside your license only.

Question keywords and the long tail

Question keywords are the most reliable source of winnable cannabis traffic, because most of them sit in the long tail tools barely see. "Can you" is especially productive here, since so much of the audience is unsure what is legal or allowed: "can you take edibles on a plane", "can you go to a dispensary without a card".

The big comparison terms are the trap. "Edibles vs smoking" belongs to Leafly, Healthline and Weedmaps, and every content calendar reaches for it anyway. Point the effort at angles nobody has written well:

• Use cases: "low thc strains for beginners", "discreet ways to consume at a festival".
• Troubleshooting: "cart not hitting", "gummies melted together".
• How-to: "how to store live resin", "how to read a COA", "how to clean a grinder".
• First-timer: "first time at a dispensary what to expect", "first dispensary visit what to bring".

A results page full of Reddit threads, Quora answers and old forum posts is the tell for an open question. Google ranks those reluctantly because nothing better exists.

Keep a question only if it has search demand or removes a purchase blocker. "Do you take debit cards" draws few searches but unblocks a sale, so it stays. Answer each question in the first sentence and add detail after. That sentence is what Google pulls into featured snippets, and what AI answers in ChatGPT, Claude, Gemini, Perplexity and Google AI tend to quote. AI search is one more surface the same pages serve.

As an example of a zero-volume question paying off: whether live resin carts go bad keeps recurring in extracts subreddits and shows no volume in any tool. A direct answer covering signs of a degraded cart, storage temperature and shelf life can rank for a dozen phrasings within months and out-earn a "best carts" listicle chasing a contested term.

Reading keyword difficulty

Keyword difficulty in cannabis is read off the results page, and tool scores only help you sort. Ahrefs KD counts referring domains pointing at the current top ten. Semrush mixes in more factors on a different scale, so one term can read 12 in one tool and 55 in the other. Neither measures content quality or the map pack. Use one tool to sort a long list, then check the shortlist by hand.

The hand check takes under a minute per term in an incognito window. If Leafly, Weedmaps, AllBud, Healthline, multi-state operators or a state .gov page hold the top five, the term waits. If position four is a 2019 forum thread and position six is a thin post with no author, the door is open. Marketplaces pile onto anything with commercial value, so a KD of 20 on a cannabis term often behaves like a 40. A term is reachable when the ranking sites are near your authority, or when their content is visibly worse than what you can write from operating experience.

Two Runtz keywords show the gap. "Runtz strain effects" reads as moderate in Ahrefs, but the SERP is Leafly, AllBud and Wikileaf, and a new dispensary blog won't crack it in a year. "Runtz clones denver" shows almost nothing in tools, and its SERP is a couple of weak marketplace listings and a stale forum post. A local page with real photos, COA data and pickup details could reach the top five in weeks.

Before writing, run three free checks: the phrase or a close variant appears in autocomplete, the SERP shows pages answering that exact question (forum scraps are even better), and Google Trends shows a line that holds or climbs over five years. A single spike that decays is a news event. New legal markets deserve speed too: their local results stay soft for a while, and the first useful page tends to hold.

Scoring the list so the order of work is obvious

Prioritizing cannabis keywords works best as simple arithmetic. Score each cluster from one to five on three axes, then multiply.

• Intent: buy, near me, deals, a city or a brand score five. Comparisons score three or four. Definitions score one or two.
• Winnability: a page one of Leafly, Weedmaps and MSOs scores one. Forums, Reddit and thin pages score five.
• Revenue proximity: a ranking category page scores five. A comparison guide linking to the menu scores three. A glossary post three hops from anything sellable scores one.

Multiplying is deliberate, because a term you can never win should sink no matter how good its intent. Picture a Denver shop scoring three terms. "Buy edibles denver" gets 5, 3 and 5, for 75. "Edibles dosage chart" gets 2, 2 and 3, for 12. "What is THCA" gets 1, 2 and 1, for 2. The order of work is settled, and re-running it quarterly shows when deferred terms become reachable.

Informational terms earn a place when they sit one step from a product. A myrcene guide with a short section linking to the strains on your menu that carry it feeds those pages daily. Before greenlighting any topic, name the product page the finished article will link to. Judge those pages on assisted conversions and branded search growth, as our guide on tracking conversions from cannabis organic search explains.

Clustering keywords and mapping them to pages

Keyword clustering groups terms by what Google shows for them, because word similarity misleads. "Edibles dosage" and "edibles recipes" look like siblings, but their results share almost nothing. "Weed delivery" and "cannabis delivery" read differently and return nearly identical results. Same results, same page. Different results, different pages.

For a small list, search both terms in an incognito window and count shared URLs in the top ten. Four or more shared results means one page can likely rank for both. For hundreds of terms, Ahrefs Parent Topic or a dedicated clustering tool does the counting.

Each cluster gets one pillar targeting the head term, and every subtopic that failed the overlap test becomes a supporting page. For edibles, the pillar is a complete edibles guide, supported by pages on dosage, onset, storage and flying with edibles. Supporting pages link up to the pillar with descriptive anchors, and the pillar links down from the section that mentions each subtopic. As an example of the math, 38 edibles terms might collapse to nine pages once overlap is checked. The supporting pages usually rank first and carry the pillar up.

Write the map down in six columns: keyword, intent, target URL, priority, status and notes, with variant rows grouped so they visibly share one URL. A seed like "wedding cake strain" can expand to forty keywords and still need only three or four pages. Menu categories work the same way: one crawlable, stable URL per query family (/menu/flower, /menu/edibles, /menu/vapes) with 150 to 250 words of real copy above the grid. If your menu is a standard iframe embed, the products inside count for the platform's domain, and our answer on Dutchie menu SEO explains the fix. Then link downhill, explainer to comparison to category, so a reader landing at the top is two clicks from your menu.

Finding and fixing keyword cannibalization

Keyword cannibalization is two or more of your pages ranking for the same query and trading places, splitting every signal you earn for that term. The check takes five minutes in Search Console: open the Performance report, click a query, then open the Pages tab. Two or more URLs with impressions means Google can't pick a winner. Position flicker, 6 on Tuesday and 14 on Friday, is the second signal.

A typical dispensary version: a Blue Dream product page, a Complete Guide to Blue Dream post and a category page, all with the strain in their titles, splitting "blue dream strain" with none above position 8. The fix gives the product page the buying intent, retitles the post around effects and grow questions, and pulls the strain name from the category title. Three fixes cover most cases:

• Merge. Fold the weaker page into the stronger one with a 301 redirect, which carries its links and history.
• Differentiate. Retarget one page at a clearly different intent, like "buy blue dream" versus "blue dream effects".
• De-optimize. For pages that must exist, remove the competing term from the title and H1.

Deleting a page without a redirect always loses. Product lines collide too. Live resin is both a dab concentrate and cart oil, so assign the plain term to concentrates, give vapes "live resin carts", and link the two pages to each other. Most cannibalization comes from a writer covering a topic the site already ranks for, so checking the keyword map comes before anyone writes.

Seed bank and clone keywords

Seed bank and clone keywords serve three different buyers, and both the list and the site should keep them apart. A seed buyer plans ahead and compares pack sizes, breeders and prices. A clone buyer wants a freshly rooted cutting of verified genetics and cares how it ships. A teen buyer wants an established plant that skips weeks of vegetative growth, usually a commercial grower on a schedule.

Grow-trait modifiers do the sorting. Feminized searchers want to avoid males, autoflower searchers are often newer and shop on ease, and photoperiod searchers skew experienced. Build those words into category pages. Most demand follows strain plus format: "blue dream seeds", "gorilla glue clones for sale", "wedding cake feminized seeds". Build a page per combination you actually stock, and split formats. A Blue Dream seeds page covers pack sizes, breeder and germination policy, while the clone page covers rooting status, ship days and replacement policy. A merged page ranks for neither, because it can't decide which buyer it serves.

Shipping and legality questions decide the sale: can clones ship to my state, will they survive the box, is it legal to buy seeds where I live. They are high-intent and weakly contested, because most seed banks bury them in an FAQ. Give each a plain page stating what your store does and does not do, without legal advice, and use breeders' published numbers for flowering time and yield.

Using Search Console as your keyword source

Search Console holds the best cannabis keyword list you will ever get: real queries, impressions and positions. Open the Performance report and set the date range to the full 16 months, because 4/20, harvest season and the spring grow rush hide inside a three-month view.

Positions 8 to 20. These are queries where Google already finds your page relevant and trusts someone else's more. Improving a page at position 12 is a fraction of the work of ranking a new one: add the missing section, answer what the title promises, earn an internal link. Filter for buy, delivery, near or price to find the commercial ones.

Regex filters. The Custom (regex) query filter splits the report. A question pattern like ^(how|what|why|can|is|do|does) pulls FAQ material. A theme pattern like gummies|edibles|cartridge|clone shows each category as a group, and a theme with impressions across dozens of queries but no dedicated page is a gap with proof attached.

Branded and generic. Branded queries click through at several times the generic rate, so blended numbers flatter you. Build one case-insensitive regex for your name and its misspellings and report the two lines separately.

Impressions against clicks. Thousands of impressions with few clicks is a packaging problem. Rewrite the title to answer the query.

To illustrate: picture "live resin vs distillate" at position 11 with about 2,000 impressions over the window and a handful of clicks, all landing on a general concentrates page. One dedicated comparison guide, linked from that page and two related posts, gives the query a real owner and a path into the top five. Remember that position is an average across locations, so pair Search Console with a geo-grid tool for map pack terms.

Keeping the keyword list current

A cannabis keyword list decays without upkeep, so build two habits around it. Monthly, sort the query report by impressions and add anything real that's missing from the map. Quarterly, compare the last six months against the six before and sort by click loss. Check each decayed query year over year before reacting, since a dip that holds against last year is a season. Retire targets that died honestly, build pages for rising queries with none, and re-check terms you once scored unwinnable.

Then put the seasons on a calendar. 4/20, 7/10 for concentrates, Green Wednesday, the spring rush for seeds and clones, and Croptober all ramp weeks before the date. Find where the Trends line leaves the floor, publish ahead of it, and keep one URL per event that you refresh yearly so its history compounds.

If you want an outside read on where your pages stand against the terms that matter, our cannabis SEO audit maps your keywords, flags cannibalization and ranks the fixes by revenue.

Frequently asked questions

Can I use Google Keyword Planner for cannabis keywords?

It makes a poor primary source. Google Ads restricts THC promotion, so the planner sees little advertiser activity on cannabis terms and its data is thin. Use Ahrefs or Semrush for estimates, and lean on autocomplete, People Also Ask and your own Search Console data.

Why do keyword tools show zero volume for cannabis terms people clearly search?

Tools estimate volume from clickstream samples and ad data, and cannabis gets weak coverage from both. Check autocomplete, the live results page, Google Trends and your own Search Console impressions before deciding a term has no searchers.

How do I know if two cannabis keywords belong on the same page?

Search both in an incognito window and count the shared URLs in the top ten. Four or more shared results is a strong sign one page can rank for both. Mostly different results mean each term deserves its own page.

Should a dispensary target "dispensary near me" on its website?

Not as words on a page. Google reads the searcher's location and ranks nearby businesses on relevance, distance and prominence. A complete Google Business Profile, exact hours, steady reviews and a real location page win near-me searches.

How often should cannabis keyword research be updated?

Add new queries from Search Console monthly and run a fuller refresh each quarter, checking year over year for seasonality. Seasonal pages for dates like 4/20 and 7/10 should be refreshed weeks before each ramp starts.