How Do You Measure Dispensary Customer Acquisition Without Reliable Ad Platform Tracking?

A customer sees a sponsorship on Thursday, searches for the dispensary on Saturday, checks the menu from a different phone, and pays cash in the store. Which platform gets credit?

Probably none of them can prove the whole trip. And that is where dispensary reporting often goes wrong: the dashboard with the cleanest number is treated as the truth, while the messy path that actually brought the customer in disappears.

Your goal is not to prove the origin of every sale beyond doubt. Your goal is to collect enough honest evidence to decide where the next marketing dollar has the best chance of producing a profitable new customer.

Start at the cash register and work backward

A click is not a customer, and neither is a video view, menu visit, or loyalty signup. Those actions can be useful steps, but acquisition occurs when the outcome you decided to buy actually happens.

For many dispensaries, the cleanest working definition is a person making a first completed purchase at any location owned by the business during a stated lookback period. You may need a different definition, but you should choose it before comparing channels.

Decide how you will handle anonymous cash buyers, merged profiles, delivery orders, people returning after a long absence, and customers whose first purchase occurred at another location. If one report counts signups and another counts first purchases, their acquisition costs are not comparable no matter how polished the chart looks.

Think in evidence, not credit

Attribution is more useful when you treat it as a confidence ladder:

• Confirmed: a unique campaign code or controlled redemption connects directly to a qualified first purchase.
• Probable: a person follows a tagged campaign link, gives permission to continue the relationship, and later completes a first purchase inside a reasonable window.
• Directional: branded searches, direct visits, customer recall, foot traffic, or store sales rise while the campaign is active.
• Unknown: the purchase happened, but the available evidence does not support assigning it to a source.

Unknown is not a reporting failure. It is an honest category that prevents weak evidence from being promoted into certainty.

Give campaigns receipts you control

You do not need one magical tracking system. You need a consistent trail of small receipts that can survive platform limits, device changes, privacy choices, and an offline purchase.

Use dedicated landing pages, QR destinations, campaign IDs, permitted offer codes, event registration fields, and an optional point-of-sale source question. Google’s official campaign URL guidance explains how UTM parameters identify the source, medium, and campaign that referred a visit.

Choose a naming convention once and use it across the link, landing page, customer system, point of sale, and reporting sheet. `spring_launch`, `SpringLaunch`, and `spring-launch` may look equivalent to a person, but inconsistent names can fragment one campaign into several rows.

And keep personal information out of campaign URLs. A source code needs to identify the marketing effort, not the customer. Names, email addresses, phone numbers, purchases, or health-related information do not belong in tracking parameters.

Calculate what acquisition really cost

The media invoice is only one part of acquisition cost. Include creative production, agency or contractor work, landing-page and tracking costs, acquisition-only discounts, sponsorship or event expenses, and staff time that would not otherwise have been spent.

The basic calculation is:

Customer acquisition cost = acquisition spend / attributed new customers

But the confidence ladder changes how you discuss the result. Suppose a campaign costs $6,000 and produces 30 confirmed new customers plus 20 probable ones. The confirmed-only acquisition cost is $200. If all 20 probable customers are included, it is $120.

The honest conclusion is not that the campaign has a precise $120 CAC. It is that the available evidence places acquisition cost between $120 and $200, depending on how much confidence you assign to the probable group. That range is far more useful than a false exact number.

A first sale can still be an expensive customer

A $100 purchase does not provide $100 to recover the marketing cost. Product cost, transaction expenses, discounts, and other variable costs have to be removed before you know what the purchase contributed.

Then watch what happens next. A channel with a higher initial acquisition cost may be the better investment if its customers return sooner, make profitable repeat purchases, and remain engaged without requiring another expensive discount.

You should consider tracking first-purchase contribution, the percentage making a second purchase, days to that purchase, and contribution over 30, 60, or 90 days. Use observed customer groups, and label any lifetime-value forecast as a forecast rather than presenting a few optimistic weeks as settled behavior.

Do not confuse timing with lift

If sales rise during a campaign, the campaign may have helped. It may also have been payday, a holiday, better weather, longer store hours, new inventory, a price change, or three other promotions running at the same time.

Compare performance with a relevant baseline by location, weekday, and season, and record what else changed. When possible, stagger the campaign across reasonably comparable stores or audiences. Local retail rarely gives you a perfect experiment, but even an imperfect comparison is better than crediting every change to the newest ad.

Collect less customer data, but make it more useful

Cannabis purchase information can be sensitive even when a particular law does not give it a special label. You should use the minimum information needed, restrict access, document retention, and make sure the business can honor applicable deletion and opt-out requests.

Do not upload a raw customer list to an advertising platform merely because the interface offers that option. First confirm the business has the necessary permission, understands the platform’s terms, and can justify the privacy risk. Better attribution is not automatically worth broader exposure of customer information.

End every report with a decision

A useful report separates confirmed, probable, directional, and unknown results. It shows total spend, acquisition-cost range, first-purchase contribution, early repeat behavior, and the assumptions behind the numbers.

Then it answers one practical question: should you stop, change, continue, or scale the campaign?

Good attribution does not make uncertainty disappear. It keeps uncertainty from making the decision for you. Once you can see which evidence is strong, which is merely suggestive, and what remains unknown, you can place the next marketing dollar with far more confidence than any single platform dashboard can provide.