# How to Choose a Cannabis Marketing Agency (and the Pitches to Walk Away From)

**Author:** John Morabito (Founder, /winston)
**Published:** September 20, 2026
**Reading time:** 12 minutes
**Canonical:** https://www.winstondigitalmarketing.com/playbooks/how-to-choose-a-cannabis-marketing-agency/

**Short answer:** Three pitches should end the meeting: anyone offering to run Google or Meta ads for THC products, anyone promising a specific number of leads, and anyone who will not name the cannabis channels they have personally worked in. What survives those three is a short list, and you sort it by asking four operational questions about your menu, your age gate, your Weedmaps listing, and your Google Business Profile. The people who have done the work answer with mechanics. The resellers answer with adjectives.

If you hold a license, your inbox is a queue. Some of it is from people who have genuinely run marketing for dispensaries, and most of it is from people who bought a list of new license holders and are pitching the same deck they send to dentists with the word cannabis swapped in. The two are hard to tell apart from the outside, because both of them sound confident and both of them talk about growth.

The good news is that cannabis makes the sorting easier than it is in most categories. The channel restrictions that make your marketing hard also make bad pitches obvious, because a pitch that ignores them is a pitch from somebody who has never run this work. You do not need to evaluate anyone's strategy. You need to check whether they know what is closed to you, and then whether they have opinions about the things that are open.

General information, not legal advice. Advertising rules, license conditions, and what a marketing vendor may do on your behalf vary by state, by license type, and by what you sell, and they change. Nothing here tells you what your jurisdiction requires or what a specific contract means for your license. Take those questions to your own counsel and your state regulator. What follows is about how to evaluate a vendor, which is a business decision rather than a legal one.

## Three pitches that end the conversation

Each of these is disqualifying on its own. You do not need a second data point and you do not need to be polite about it.

### 1. Anyone offering to run Google Ads or Meta ads for THC products

Both platforms prohibit it, and the prohibition holds in states where what you sell is perfectly legal. This is not a gray area that a good rep can talk you through (https://www.winstondigitalmarketing.com/playbooks/can-you-advertise-cannabis-on-google/), and it is the single fastest test of whether the person across from you has worked in the category at all.

There are two versions of this pitch and both are bad. The first is honest ignorance: they sell paid search to every other vertical, nobody told them cannabis is different, and they will discover the problem after you have signed. The second is worse and more common than you would like. They know, and the plan is a workaround: a landing page that does not mention what the store sells, an account set up under a different entity, a campaign that runs for a few weeks before something catches it. That puts your account, your domain, and sometimes your business relationships on the line so that an agency can show activity in month one.

The nuance worth allowing for is that some adjacent and ancillary advertising is genuinely permitted in narrow circumstances, and the rules are not identical for every product or every platform surface. Someone who knows the category will make that distinction themselves, unprompted, and will be specific about which product and which policy. Someone who waves at it as a technicality they have handled before is telling you exactly what they intend to do.

### 2. Anyone promising a fixed number of leads

Forty qualified leads a month. Fifteen new customers a week. A guaranteed return on the retainer. In a category where the ad platforms are closed, the menu often lives on somebody else's domain, and a large share of purchases happen in a store with no digital record connecting them to anything, nobody can promise that number, because nobody can even measure it cleanly after the fact.

Which is the actual problem with the promise. Ambition is fine. The trouble is that the only way to be held to a number like that is to define a lead loosely enough that it can be produced on demand, and then you spend month four arguing about whether a form fill from somebody three states away counts. By that point the guarantee has stopped working as a commitment and started working as a way to make the invoice defensible.

An agency that is confident and honest at the same time will tell you what it expects to move and over what period, will name the leading indicators it expects to move first, and will say plainly which of them it does not control. That is a harder sentence to say in a sales meeting than a guarantee is, which is part of why it is worth something.

### 3. Anyone who will not say which channels they have actually run

Ask directly: which of our channels have you personally worked in, on which accounts, and what did you do there. Not which industries the company serves. Not which logos are on the website. Which channels, which accounts, which work.

The answer sorts people quickly. An operator says something like: I ran Weedmaps and the Google Business Profile for two shops in Michigan, I have never touched a Dutchie menu, and here is what I would want to learn in the first month. That answer contains a gap and is more credible for it. A reseller answers with a capability list, then routes the question to a case study with no names in it, then offers to bring in a specialist on the next call. The specialist, if they exist, is frequently a subcontractor you are about to pay a markup for.

None of this is unique to cannabis. Every category has resellers. What is specific here is that the cost of hiring one is higher, because the channels you are left with are hands-on and unglamorous, and a vendor who is one layer removed from the work will quietly deprioritize all of them in favor of the deliverables that are easy to produce in volume. The same test, asked slightly differently, is the one I recommend for choosing a GEO agency (https://www.winstondigitalmarketing.com/playbooks/how-to-choose-a-geo-agency/), where the failure mode is identical and the vocabulary is newer.

## Four questions that separate an operator from a reseller

Once the disqualifiers are out of the way, stop asking about strategy. Strategy questions are easy to answer well and tell you almost nothing. Ask about four specific pieces of your own infrastructure instead, and listen for whether the answer has mechanics in it.

### What would you do about our menu?

Most dispensary menus are embedded from a third party, which means the products, the prices, and the strain names live on somebody else's domain and never appear in your own HTML. The inventory that people are actually searching for is invisible to search engines and to AI assistants, and the page that should be your strongest commercial asset is a wrapper around a frame.

A good answer engages with that: whether the provider offers a rendered or indexable option, whether you can build your own product pages alongside the embed and keep them in sync, what you lose either way, and how much of it is worth doing given what your menu turnover looks like. A weak answer treats the menu as a design element. The full version of this problem, including which fixes actually work, is in the Dutchie iframe SEO problem (https://www.winstondigitalmarketing.com/playbooks/dutchie-iframe-seo-problem/) and dispensary menu sync (https://www.winstondigitalmarketing.com/playbooks/dispensary-menu-sync-seo/).

### What would you do about our age gate?

Every licensed operator has one and most of them are built in a way that serves a crawler the door instead of the page. Ask what they would check first. The answer you want involves finding out what the server actually returns to a plain request, because there are three common implementations and only one of them leaves a readable page behind. That diagnosis is a five minute job and it is covered in the age gate playbook (https://www.winstondigitalmarketing.com/playbooks/cannabis-age-gate-seo/).

If the response is that the age gate is a compliance matter outside marketing, you have learned something useful. It is the single most likely reason a cannabis site is invisible, and an agency that does not know that will sell you content for a year that lands behind a closed door.

### Who updates our Weedmaps listing, and how often?

Weedmaps and Leafly are where a large share of high-intent comparison happens, and the listings are operational rather than strategic: hours, menu accuracy, deals, photos, responses, the specials that are still showing after they ended. Someone who has run this will have an opinion about cadence and about which fields matter, and will ask who at your shop currently owns it, because the answer is usually a manager doing it between shifts.

Someone who has not will describe it as listing management and move on. The tell is whether they treat it as a recurring operational task with an owner and a frequency, or as a line item. The detailed version is Weedmaps optimization (https://www.winstondigitalmarketing.com/playbooks/weedmaps-optimization/).

### What is the first thing you would change on our Google Business Profile?

Have it open during the call. This one is unfair in a useful way, because a person who has done local work for retail will look at the profile and immediately name something concrete: a category that is wrong, hours that do not match the site, a products section that was never filled in, a review backlog, photos from the buildout. It takes them ten seconds and it is not a rehearsed answer.

Someone who has never run a profile will talk about optimizing it. That word is doing all the work in that sentence, and there is nothing underneath it. The dispensary Google Business Profile playbook (https://www.winstondigitalmarketing.com/playbooks/dispensary-google-business-profile/) covers what the concrete version looks like.

## What an honest scope looks like when paid search is closed to you

When you cannot buy attention, the work stops being a media plan with support around it and becomes six or seven ongoing operational jobs that compound slowly. That is a genuinely different shape of engagement, and a scope that has not been rebuilt around it will read like a normal marketing retainer with the ads line quietly deleted.

An honest cannabis scope usually contains most of these, with a clear owner and a clear cadence for each:

- **The site, as a technical asset first.** The gate, the menu, the speed, whether pages are actually reachable and readable. This comes before content, always, and it finishes, which means it should not sit on the invoice forever.
- **Local search and the profile.** The map pack is the highest-intent surface you control, and the profile needs attention every month.
- **The cannabis-native platforms.** Weedmaps, Leafly, and whatever your market uses. Operational, weekly, unglamorous.
- **Content that is built to be cited.** The one channel with no ad policy ceiling, which is the argument in cannabis content marketing (https://www.winstondigitalmarketing.com/playbooks/cannabis-content-marketing/), and increasingly the way you show up in AI answers rather than only in blue links.
- **Owned lists.** Email and SMS, where the compliance requirements are real and the returns are the most reliable in the category.
- **Reviews and reputation.** Load-bearing for local search and for AI answers both, handled as a routine rather than a campaign.

Two things to look for in how that scope is written. First, whether it is small enough to actually finish every month. A list of twelve deliverables at a modest retainer is a list of twelve things that will each get an hour. Second, whether anything in it is sequenced. If the site is broken, the content work should be scheduled after the fix, and a scope that runs all six workstreams from day one is a scope designed for invoicing rather than for results.

Ask one more question about the scope: what gets cut first if the budget drops by a third. The answer tells you what they think is load-bearing, and it is a much better question than asking what they would do with more money.

## How to price it against the size of the shop

The right retainer is a function of your business, not of the agency's package tiers. A single store in one neighborhood, a three-location operator, and a brand with a wholesale arm selling into other people's shelves are three different companies with three different marketing problems, and any agency that quotes them the same number is selling a template and hoping.

Rather than benchmarking against a number somebody quoted in a forum, reason it through from your own side:

1. **What is the marketing actually supposed to influence?** For a single store that is foot traffic and repeat visits in a defined radius, which is a bounded number you can estimate from your own sales. The fee should be a defensible share of that, not of some category-wide figure.
2. **How much of the work is a one time fix versus a recurring job?** Technical site work, a menu strategy, and a profile rebuild are projects. Listings, content, lists, and reviews are ongoing. Paying a permanent retainer for work that finishes in two months is a common and avoidable mistake, and a good agency will tell you which parts those are.
3. **What does the same money buy in house?** At some scale, a capable person on staff plus a specialist for the technical work beats a retainer. Knowing where that line sits for you is useful leverage, and an agency worth hiring will discuss it honestly rather than flinch.
4. **Can you fund it for long enough to matter?** The channels open to cannabis compound over quarters. A retainer you can sustain for a year at a smaller number beats a larger one you cancel in month four, which produces nothing at all. If the number is only affordable in a good month, it is the wrong number.

The allocation question underneath this one, meaning how to split whatever you have across the channels rather than how much to spend in total, is the whole subject of the cannabis marketing budget playbook (https://www.winstondigitalmarketing.com/playbooks/cannabis-marketing-budget/). Read that before the pricing conversation, because it is hard to evaluate a proposal when you have not decided what the money is for.

## What reporting should contain when most platforms will not attribute

This is where a lot of cannabis engagements quietly go bad. The category has a measurement gap that nobody can close: purchases happen in store, the menu is on another domain, the ad platforms that would normally stitch the path together are not available to you, and a meaningful share of demand arrives through directories that report only what they choose to report.

Agencies respond to that gap in one of two ways. The honest response is to be explicit about the line between what was measured and what is inferred. The other response is to fill the gap with a confident number, usually a revenue figure or a return multiple, produced by a model with assumptions nobody in the meeting has read. The second kind of report is more comfortable to receive and it is worth less than nothing, because decisions get made on it.

**What belongs in the measured column:** impressions, clicks, and query data from Search Console; calls, direction requests, and profile views from the Google Business Profile; menu views and orders where your menu provider reports them; list size, delivery, and click rates on email and SMS; review volume and rating; redemptions on a code or an offer that only exists in one channel; and rankings, with the caveat that rankings are an input rather than an outcome.

**What belongs in the inferred column, clearly labeled:** anything connecting an impression to a purchase through a path no platform observed end to end. Inference is fine as long as it carries the label. A report that says we believe this content is contributing to in store demand, here is the indirect evidence, and here is what we cannot see, is a more useful document than one that reports a precise revenue number nobody can reproduce.

Two things worth asking for specifically. One is a recurring check on whether you are appearing in AI assistant answers for the obvious category questions in your city, because that is where a growing share of the discovery you cannot measure is happening, and it is checkable by hand even though it is not reportable in a dashboard. The other is a short written narrative each month: what we did, what moved, what did not, what we are doing next. In a category this hard to attribute, the narrative is frequently the only part of the report that carries real information.

## Contract terms worth negotiating when a license can change

Cannabis businesses carry a category of risk that most agency contracts are not written for. A license can be suspended, a renewal can stall, a municipality can change its position, an ownership change can be required, a location can be blocked after a lease is signed. Standard agency paper assumes the client will keep existing in roughly the same form for the whole term, and that assumption is doing a lot of quiet work.

Terms worth raising before you sign, all of which are ordinary business points rather than exotic asks:

- **A short initial term, or an out that is not punitive.** Twelve months with no exit is a long time to be wrong about a vendor. Ninety days, then month to month, is a reasonable ask and a reasonable thing for an agency to accept if it believes it will earn the renewal.
- **A suspension clause tied to license status.** If your license is suspended or your operations are paused by something outside your control, you want the ability to pause the engagement rather than to keep paying for work that has nowhere to land, or to breach.
- **Ownership of everything produced, stated plainly.** The site, the content, the accounts, the listings, the profile, the list. Ownership should sit with you on creation, not on final payment, and definitely not with the agency.
- **Access in your own name.** Your Google Business Profile, your Search Console, your analytics, your email and SMS platform, your Weedmaps account, all owned by an entity you control with the agency added as a user. An agency that holds your accounts holds your business, and this is the term most often skipped and most painful to fix later.
- **An offboarding obligation.** What is handed over, in what format, within how many days. Write it down while everyone likes each other.
- **Who is responsible for compliance review.** An agency can and should know the platform rules cold. It is not your regulatory counsel, and a contract that implies otherwise protects nobody. Be clear about who reviews what before it goes live.
- **Subcontracting, disclosed.** If parts of the work go to another firm, you are entitled to know which parts, and to know whether your accounts are being accessed by people whose names you have never seen.

Have counsel look at the agreement. That sentence is in every article like this one and it usually gets skipped, so here is the specific reason it matters more in this category: the terms above interact with your license conditions in ways that differ by state, and the clause that is boilerplate in one jurisdiction can be a problem in another.

## What to do this week

1. Take the three disqualifiers to whatever is sitting in your inbox right now. Expect most of it to fail at the first one.
2. Ask the four operational questions on the next call, with your Google Business Profile open in front of you.
3. Decide what the money is for before you evaluate what it costs. The allocation framework is in the budget playbook (https://www.winstondigitalmarketing.com/playbooks/cannabis-marketing-budget/) and the sequencing is in the dispensary marketing plan (https://www.winstondigitalmarketing.com/playbooks/dispensary-marketing-plan/).
4. Ask any finalist what gets cut first if the budget drops by a third, and what they would not do at all.
5. Ask for a sample report from a live account with the names removed, and check whether measured and inferred are separated anywhere on it.
6. Send the agreement to counsel with the account ownership and license suspension terms flagged specifically.

The category rewards hiring slowly. The channels that work in cannabis compound over quarters, which means a wrong choice costs you a year rather than a month, and the pitches designed to produce a fast yes are the ones selling the things that do not work. If you want a second read on a proposal in front of you, that is the kind of thing we look at as part of cannabis marketing (https://www.winstondigitalmarketing.com/services/cannabis-marketing/), including when the answer is that the proposal is fine and you should sign it.

## Frequently asked questions

### What does a cannabis marketing agency actually do if it cannot run ads?

It works on the channels that are open to you, which is most of them. That means the website and its technical health, local search and the Google Business Profile, the cannabis-native platforms where high-intent shoppers actually compare menus, content that can be read by both search engines and AI assistants, owned lists like email and SMS, and reviews. An agency that cannot describe that work without reaching for a paid media budget has not adapted its model to your category. The absence of paid search is not a smaller version of a normal engagement. It changes which channel carries the load, which is why the scope should look different from the one they would sell a restaurant.

### Can a cannabis marketing agency run Google Ads or Facebook ads for THC products?

No, and an agency offering it is either uninformed or planning to work around the rules with your account and your license as the collateral. Both platforms prohibit advertising for recreational and unapproved cannabis products, and that prohibition holds in states where cannabis is legal. The workarounds that exist all involve misrepresenting what the account is selling, which risks the account, the domain, and the business relationship, and none of it is worth a click. Some adjacent and ancillary advertising is genuinely permitted in narrow cases, so the honest version of this conversation is specific about which product and which policy. Confirm anything you are told with the platform policy itself and with your own counsel.

### What questions should I ask a cannabis marketing agency before signing?

Ask four operational questions and listen for whether the answers are specific. What would you do about our menu, given that it is embedded from a third party. What would you do about our age gate. Who updates our Weedmaps listing and how often. What is the first thing you would change on our Google Business Profile. Someone who has run this work answers with the mechanics and the tradeoffs. A reseller answers with a benefit statement. Then ask who does the work, whether any of it is subcontracted, and which cannabis accounts they have personally run rather than which ones their company has logos for.

### How much should a dispensary pay a marketing agency?

Size it against the shop rather than against a package. A single location paying rent in one neighborhood cannot support the same retainer as a multi-location operator with a wholesale arm, and an agency that quotes both the same number is selling a template. The useful test is whether the fee is a defensible share of what the marketing is supposed to influence, and whether the scope inside it is small enough to actually get done every month. A modest retainer spent entirely on the owned channels usually beats a larger one spread across six deliverables nobody finishes. Ask what gets cut first if the budget drops by a third, because the answer tells you what they consider the load-bearing work.

### What should cannabis marketing reporting include when platforms will not attribute?

It should separate what is measured from what is inferred, and say which is which on the page. Measured includes rankings and impressions from Search Console, calls and direction requests from the Google Business Profile, menu views and orders where the menu provider reports them, list growth and campaign performance on email and SMS, and store visits or redemptions tied to a code. Inferred includes anything connecting an impression to a sale through a path that no platform observed end to end. A report that presents a single confident revenue attribution number in a category where most of the chain is unobservable is telling you a story. Ask for the leading indicators and the limits stated plainly instead.

## Read these before the pricing conversation

- Cannabis marketing budget: what to spend and where: https://www.winstondigitalmarketing.com/playbooks/cannabis-marketing-budget/
- The dispensary marketing plan: https://www.winstondigitalmarketing.com/playbooks/dispensary-marketing-plan/
- Can you advertise cannabis on Google?: https://www.winstondigitalmarketing.com/playbooks/can-you-advertise-cannabis-on-google/
- How to choose a GEO agency: https://www.winstondigitalmarketing.com/playbooks/how-to-choose-a-geo-agency/

## Want a second read on a proposal?

We will look at what is in front of you and tell you whether it is worth signing, including when it is. This page is general information, not legal advice. Confirm advertising rules and contract questions with counsel and your state regulator.

Service: https://www.winstondigitalmarketing.com/services/cannabis-marketing/
Audit: https://www.winstondigitalmarketing.com/audit/