cannabis · regulated markets
Why Dispensary Influencer ROI Disappoints 9 in 10 Brands
Mayra runs a Toronto dispensary on Bloor Street. We pitched her on creator deals like the ones CBII CBD, a UK-based CBD brand, ran 9 times with Jamie Genevieve, a 1.4M-subscriber Scottish beauty creator, between June and December 2023. CBD is cannabidiol, the non-intoxicating part of hemp. THC gets you high. Mayra wanted the same setup. The answer was no.
A hemp brand ships to all 50 states. A dispensary ships to one. Same creator, same post, one tenth of the buyer pool. That gap kills dispensary ROI.
Across the 23 cannabis brands and 66 cannabis YouTube channels we track, zero of the named hemp anchors run a single-state retainer pattern.
The geography lock
A hemp brand pays for a creator's whole audience. So does a dispensary. Only the hemp brand can sell to the whole audience.
Cornbread Hemp, a Kentucky hemp-derived CBD brand, ran 7 deals across 6 creators between July 2025 and January 2026. Hemp ships to every state under the 2018 Farm Bill, legal at less than 0.3% THC. Every viewer is a buyer. A dispensary pays the same fee, but only viewers in one state can walk in.
The brand pays for 100% of the reach. It can address 8 to 15% of it. That is the lock tax.
Here is the in-state audience pull we run before any dispensary deal goes out.
The in-state audience math
A Toronto dispensary pays a fee for a cross-state cannabis creator. Maybe 6% of that audience lives in Ontario. The cost per addressable viewer is the fee divided by 6%, not by the full audience. Run it against a Toronto creator where 60% of viewers live in the city, and the per-viewer cost falls by ten times.
The same math broke our Toke Cannabis pitch. Here is what we wrote Mayra:
Torontian here and love your cannabis store, always pass by it at Bloor St. I work with thousands of top Toronto influencers that make locations in downtown go viral, like Crumbl Cookies.
She read it. She did not reply. The pitch should have led with the geography filter. Here is the Toronto-only audience cut we send dispensaries now.
We hedge to cadence, not posted rates. None of the 12 cannabis-niche channels we track have a captured per-post rate.
Wondering if any creator clears your in-state ratio? The audit step is what we run before any dispensary deal lands. A 3-creator shortlist with audience-region check, geo-ratio, and comp-deal structure already drafted, in 14 to 21 days.
Talk to us about your store →Four ROI band archetypes
Four creator types come up over and over. Only the first three pencil for a dispensary.
The first is the dispensary-staff creator. A budtender on TikTok or Instagram. The audience self-selected for that store. The deal is a small hourly add or a sales bonus, not a flat fee.
The second is the hyper-local lifestyle creator. A Toronto food creator, an LA wellness creator, a Denver music creator. In-state audience above 60%. The geography ratio brings the per-viewer cost in line with what a hemp brand pays nationally.
The third is the niche podcast. The Randall Carlson, a 608K-subscriber Earth-history podcast host, ran 6 deals with CBD From The Gods between January and March 2026. CBD From The Gods is a small US CBD brand tied to the show. The repeating call to action was "RANDALL'S VIEWERS GET FREE SHIPPING FOR LIFE: Use Code RCSHIPSFREE." That is a hemp brand, but a dispensary with a podcast whose listeners cluster in one city can copy the structure.
The fourth is the cross-state cannabis-niche archetype. Leafly at 310K subs, CannaCribs at 332K, The Cannabis Experts at 416K, RespectMyRegion at 71K. These channels work for hemp brands. They do not work for a dispensary at any flat fee we have seen.
- Flat-fee deals burning 80% of audience reach on non-buyers
- Creators picked by follower count, not by in-state share
- No comp plus revenue-share fallback when the flat-fee math fails
one thing we hear all the time is how difficult it is running campaigns in this space with all the compliance and regulatory issues, content gets taken down, ads get restricted, and a lot of creators just don't know how to stay within guidelines while still making content that converts.— Cornbread Hemp outbound thread, April 2026See if the math pencils for your store →
Foot-traffic vs online attribution split
Half the impact shows up online. Half shows up as foot traffic. Most stores measure neither cleanly.
The cleanest mechanic we have logged is the coupon code. Sunset Lake CBD ran 4 deals across 2 creators, including The Majority Report podcast with codes "FRIDAY25" and "420". Each code tracks one creator. The brand can split online versus in-store if the code works at both registers.
Without a unique code per creator, the math is unfalsifiable. That hits dispensaries harder because the per-customer cost is already five to ten times higher.
Run the code for 30 days. Count online redemptions. Pull store POS for the same window. Compare against the prior 30 days. If the lift beats the deal cost, the post worked.
Comp-only deals that work
The dispensary that cannot pay a flat fee can still get a creator post. The structure is product comp plus a percentage of new-customer revenue.
Three things shift. The creator carries some risk. The brand stops paying the lock tax. The deal renews only if it works.
Dispensary-staff creators almost always accept comp. Hyper-local creators sometimes. Cross-state cannabis-niche creators almost never. A creator who refuses comp is telling you their fee assumes flat-fee math. For a dispensary, that math fails.
For the contrast case where a brand can ship nationwide, see our cannabis D2C vs retail creators and cannabis creator rate card breakdowns.
Where we come in
We do the in-state audience pull and the comp-deal draft before any dispensary makes its first offer. The hemp creator universe we track sits next to the dispensary-fit shortlist, so we already know which names take comp and which only quote flat fees. We also keep the FTC and FDA risk file on every name, because the rules on cannabis claims change every quarter. The FDA's position is in their explainer on cannabis-derived products. Platform rules sit in Meta's drugs and pharmaceuticals ad policy and TikTok's healthcare and pharmaceuticals policy.
Send us your store and we will run the geography read free, no pitch.
Geography decides everything.
FAQ
Can a single-state dispensary break even on influencer marketing?
Yes, under specific conditions. The creator's in-state audience share needs to clear about 35%. The deal needs to pay in product comp plus a percent of new-customer revenue. Flat fees on national cannabis-niche creators carry the full lock tax.
Why do most dispensary creator deals lose money?
The geography lock. A dispensary pays for 100% of a creator's audience but can sell to maybe 8 to 15% of it. The per-customer cost runs five to ten times what a hemp brand pays for the same post.
What kind of creator works for a dispensary?
Three archetypes. Hyper-local creators where in-state audience clears 60% or more. Dispensary-staff creators whose audience self-selected for the store. Small niche podcasts where the geography ratio lets flat-fee math pencil.
Reading loop
Frequently asked
Can a single-state dispensary break even on influencer marketing?
Yes, but only when the creator's in-state audience share clears about 35%. The deal also needs to pay in product comp plus a percent of new-customer revenue. Flat fees carry the full geography-lock tax. Comp deals do not.
Why do most dispensary creator deals lose money?
The geography lock. A dispensary pays for 100% of a creator's audience but can sell to maybe 8 to 15% of it. The per-customer cost runs five to ten times what a hemp brand pays for the same post.
What kind of creator works for a dispensary?
Hyper-local creators where in-state audience clears 60% or more. Dispensary-staff creators whose audience self-selected for the store. Small niche podcasts where the geography ratio finally lets flat-fee math work.
Next issue, every Monday
We found the best performing creators for May 25 → May 31.Hand-picked, not the same five names.
Plus the Influencer Advisory Consultant GPT.