Most pizza marketing budgets are set by inheritance. Last year’s number, plus whatever a vendor talked someone into, minus whatever got cut when food cost rose. A budget built deliberately (with a stated split between acquisition, retention and experiment) outperforms one built by accretion, mainly because it can be argued with.
Independent pizza locations typically spend between 3% and 6% of net sales on marketing in steady state. New locations and stores in contested markets run higher: 8% to 12% is normal during an opening year or a competitive fight.
Below 2%, most shops are effectively relying on location and word of mouth, which is a strategy but not a plan. Above 10% in steady state, the spend is usually substituting for a positioning or operational problem.
| Bucket | Share | What it buys |
|---|---|---|
| Retention & lifecycle | 35–45% | Email, SMS, loyalty rewards, journeys, win-backs. Cheapest revenue per dollar in the whole budget. |
| Acquisition | 30–40% | Paid search, paid social, mail, door drops, new movers. |
| Brand & community | 10–20% | Sponsorship, fundraisers, local presence, radio. Slow, compounding, hard to attribute. |
| Experiment | 10% | Deliberately unproven. Ring-fenced so it does not get raided. |
Retention marketing is cheap, which makes it a soft target when budgets are cut and an easy thing to over-fund with discounts when they are not. Both errors are common.
The discipline: fund the mechanics of retention generously, the platform, the data, the journeys, the identification work, and fund the discounting within retention against measured incremental lift only. Rewards and offers should have to earn their budget every quarter against a holdout.
Ring-fence roughly a tenth of the budget for things that might not work, and judge it on learning rather than on return. Without a protected experiment budget, an organisation only ever repeats what it did last year, and the first genuinely new channel it tries will be one it was forced into.
Rules that keep it honest: every experiment states its success criterion before it starts, runs long enough to read, and ends on a decision: scale, kill, or retest. An experiment with no stated criterion becomes a permanent line item.
A budget with no experiment line is a plan to be exactly as good next year as you were last year.
Rank every channel by incremental return per dollar, measured against holdouts rather than platform reports. Move money down the list toward the top, but move it gradually. Most channels have a saturation point where doubling the spend does not double the return, and mail and paid social both hit it earlier than operators expect.
Quarterly is the right cadence. Monthly reallocation chases noise; annual reallocation means a year of funding something that stopped working in February.
Both. A central budget for brand, technology and programmes that scale, plus a local budget each store can spend on its own trade area. Stores know their neighbourhood, their schools and their competitors better than any central team, and a local budget is what lets them act on it.
Email plus the identification work needed to make it useful. It has the lowest cost per incremental order of any channel available to a pizza shop, it compounds, and it is the foundation everything else eventually attaches to.
With matched-market or matched-period holdouts, and by accepting a longer measurement window. Brand spend that can never be evidenced at all is faith rather than budget, but the answer is better measurement, not abandoning the category.
Keep reading
Building holdout groups into every campaign, reading the difference honestly, and accepting that most reported marketing revenue was going to happen anyway.
MeasurementWhy last-click attribution systematically misvalues restaurant channels, and a practical model for splitting credit across email, SMS, search, social and mail.
MeasurementAn annual planning rhythm built from sports fixtures, school terms, holidays and your own demand curve, so the marketing is booked before the week arrives.
ChannelsWhere linear TV still works, how connected TV changed the maths for multi-unit operators, and the store count at which broadcast starts making sense.
marketing.pizza runs all of it, every night, across every store you have.