Home/Playbook/Offer architecture and margin guardrails

Campaigns & offers

Building an offer ladder that does not eat the business

Designing a coherent set of offers with fences, stacking rules, minimum spends and a maximum discount depth, so promotions stay a tool rather than becoming the price.

Campaigns & offers4 sections3 questions answered
Campaigns & offers illustration

Most pizza shops do not have an offer strategy. They have a pile of offers accumulated over several years, each one introduced for a good reason, none of them retired, several of them stackable in combinations nobody has ever calculated. The result is a shop whose effective price is unknown and whose customers have learned never to pay the menu price.

Audit what you actually have

Before designing anything, list every live offer: menu bundles, standing coupons, third-party promotions, loyalty rewards, employee discounts, local partnerships, anything printed on anything. For each one record the discount depth, the fence, the expiry, and the share of tickets it touches.

Two things typically emerge. First, a small number of offers account for most of the discount dollars. Second, several offers exist that nobody remembers authorising and that no longer serve any purpose.

Fences are what make an offer a tool

A fence is the condition that stops an offer from applying to business you already had. Without a fence, an offer is a price cut. Every offer should have at least one.

  • Time fence: off-peak only, weekday lunch only, before 5pm.
  • Channel fence: direct online ordering only. The most valuable fence a pizza shop has, because it also converts marketplace customers.
  • Audience fence: first-time customers only, lapsed customers only, loyalty members only.
  • Basket fence: minimum spend, or a required attach.
  • Product fence: applies to items with headroom rather than to the whole menu.
  • Deferral fence: redeemable on the next order rather than this one, which is the strongest fence available.

Set a maximum depth and hold it

Decide the deepest total discount any ticket may carry, and enforce it in the ordering system rather than in a policy document. A common workable ceiling is 25–30% off the pre-tax subtotal, all sources combined.

Then define stacking explicitly. The default should be that offers do not stack, with named exceptions. It is far easier to permit a specific combination later than to discover an unintended one in the P&L three months after launch.

CombinationDefaultWhy
Deferred discount + loyalty points earnAllowedPoints earn on net spend; encourages redemption without compounding cost
Deferred discount + couponBlockedTwo forward-value instruments on one ticket; depth becomes unbounded
Loyalty reward + bundle priceBlockedBundle is already discounted; the reward should apply to full-price items
Employee discount + anythingBlockedStandard control, and the most common leakage point in independents
First-order offer + deferred issuanceAllowedDifferent timing; this is the second-order journey working as designed

Measure discount as a rate, per store, weekly

Total discount dollars divided by gross sales is your effective discount rate. Track it weekly per store on the same report as food and labour. It drifts upward silently, and it is far easier to arrest at 9% than at 17%.

Watch the distribution as well as the average. A 9% average made of most tickets at 0% and a tail at 40% is a very different business from a uniform 9%, and it usually means one fence is broken.

If more than half your tickets carry a discount, you do not have promotions. You have a price list you are embarrassed about.

Questions

What discount rate is healthy for a pizza location?

Most healthy independents run a total effective discount rate between 5% and 12% of gross sales. Above 15% sustained, the discounts are usually substituting for a pricing or positioning decision that has been avoided.

Should I match a competitor's deep discount?

Almost never head-on. Matching converts a share battle into a margin battle you both lose. Fence a targeted response to the specific customers at risk instead: that costs a fraction as much and does not reset your whole market's price expectation.

How do I retire an offer customers are used to?

Replace rather than remove. Introduce a fenced or deferred version alongside, shift volume to it over four to six weeks, then withdraw the original quietly. Abrupt removal of a familiar offer produces a visible traffic dip and a lot of avoidable complaints.

Keep reading

Related chapters

All 46 chapters

This is one chapter of the job.

marketing.pizza runs all of it, every night, across every store you have.