Set a coffee-shop ad budget using order contribution and repeat behavior

A coffee shop can sell more drinks and still make an uncomfortable advertising decision if acquisition cost exceeds the contribution those orders create. Gross sales are only one part of the calculation. Ingredients, packaging, payment fees, incremental service costs and any promotion all affect what remains.

Start with a conservative first-order model. Add repeat behavior only when the business has appropriate evidence and a clear observation period. The useful question is how much acquisition cost can be supported by the expected contribution, with uncertainty stated rather than hidden.

Define contribution for the decision

Record the average basket for the offer or customer group being evaluated. Subtract the variable costs relevant to fulfilling that basket. State whether labor is included and why.

A quiet-hour promotion with spare capacity can have different incremental costs from a campaign that requires another staff member. Keep those circumstances explicit. The result is contribution before advertising and remaining business costs, not the shop's full net profit.

Also calculate the cost of the offer. A free item or discount is part of acquisition economics even when it does not appear in the advertising platform's spend column.

Work through a hypothetical first order

Assume an advertised purchase has a $9 basket. In this hypothetical example, ingredients and packaging cost $3, payment and other variable costs cost $0.50, and incremental service cost is $1.50. Contribution before advertising is $4.

If the owner wants at least $2 left from that first order for the remaining business costs, the acquisition allowance is $2 per acquired purchasing customer. This is an owner-defined planning limit, not a universal recommended CPA.

Now suppose the campaign costs $100 and produces twenty observed offer redemptions. Advertising cost per observed redemption is $5. The offer is above the assumed first-order allowance even before considering any separate discount cost.

That does not automatically prove the campaign was unprofitable overall. It means the first-order model does not support it without additional justified value or a different business objective.

Add repeat behavior cautiously

Imagine the shop has consented first-party records showing that a comparable new-customer cohort made an average of two additional purchases within sixty days. Suppose those purchases each contributed $4 under the same cost definition.

The hypothetical cohort contribution is then $12 across three purchases before advertising and remaining costs. That is a broader horizon than the $4 first-order contribution. It must be labeled as such.

Do not assume every directions click becomes a customer who repeats three times. Retain the observed customer definition, cohort dates, sample size and incomplete observation. If only a small subset can be linked reliably, say so.

Separate prediction from collected cash

A repeat-purchase estimate is not cash already received. If the business needs to protect this week's liquidity, a sixty-day value model cannot eliminate the short-term cash constraint.

Keep first-order receipts, later observed receipts and expected future contribution in separate rows. Include refunds, promotion costs and the uncertainty in repeat behavior.

Google's About conversion values concerns representing business value. Any platform value should use a clearly chosen definition rather than disguising assumed future contribution as realized sales.

Consider capacity and substitution

A campaign can shift existing customers toward a discounted offer without adding many new purchases. It can also move demand into a busy period that already sells near capacity.

Review the time of day, item availability and service capacity. A new visit during a quiet period may have a different operational effect from one that lengthens a peak queue. The model should reflect the actual decision.

Attributed sales do not establish incremental sales. Google's About Conversion Lift describes controlled incrementality measurement, which has eligibility and data requirements.

A small shop may use a narrower operational comparison while honestly acknowledging that it is observational.

Build cautious, observed and optimistic scenarios

Use three versions of the worksheet. Change only the assumptions you are uncertain about, such as repeat count, basket or variable contribution.

For the cautious scenario, use first-order value and no assumed repeat. For the observed scenario, use a mature comparable cohort. For the optimistic scenario, clearly label the stronger assumptions and avoid treating them as the budget's guaranteed outcome.

Display each scenario as a readable card on mobile. Include the contribution horizon, acquisition allowance, actual ad spend and observed purchase evidence.

Connect the economics with campaign review

The How Much Should a Small Business Spend on Google Ads? provides broader spending context. Use it alongside the shop's contribution calculation, not instead of it.

The Google Ads Value-Based Bidding: Beyond Cost Per Lead can help review account settings when values are reliable. A value model alone does not establish that the relevant platform integration or campaign strategy is available for the shop.

Also review the Count Every vs One: How Conversions Get Tallied so a purchase and its repeated confirmation do not inflate apparent value.

Choose a practical next decision

If the observed acquisition cost exceeds the cautious allowance, investigate the offer, targeting, measurement and repeat evidence before increasing spend. If the cost appears acceptable, check capacity and whether the measurement actually captures a purchasing customer.

Keep the proposed action specific and within the owner's budget authorization. No spreadsheet can turn a weak assumption into a promised return.

Explore current AdvisorPPC plans and bring the contribution worksheet to an account review. The best next question is which part of the acquisition model the available evidence supports today.