Loyalty Campaign Management: Run Campaigns You Can Price and Prove

Loyalty campaign management is how a program team plans, sets up, pays for and proves the offers it runs on top of the everyday earn rules. The base program pays members for normal behavior. A campaign pays extra for one change in behavior, from a chosen group, inside a set window. That extra pay is real money. So each campaign needs a target, a rule set and a cost ceiling, plus a clean way to show it worked.

What is loyalty campaign management?

A loyalty campaign is a time-bound offer aimed at a defined group of members. It has a start date, an end date, an audience, a reward and a goal. Campaign management runs the whole cycle, from the brief to the readout.

Loyalty marketing campaigns sit between two neighbors. Loyalty marketing sets the strategy: whose loyalty to grow and what value to offer in return. Program design sets the always-on rules. Campaigns are the short, funded pushes that test and tune both. A program that runs campaigns with no readout is spending its points budget on guesses.

What types of loyalty campaigns do programs run?

Loyalty program campaigns come in five common types. Each one asks members for a different behavior. So each one needs its own rules and its own measure.

Bonus points events

A bonus points event raises the earn rate for a short window, such as double points on one day or in one category. Starbucks Rewards runs this as Double or Triple Star Days. Its terms give Gold members four more Double Star Days a year and Reserve members six. For Green members, Starbucks picks a personal day. Members can be asked to activate a Star Day before it counts (Starbucks).

Each of those rules narrows who collects the bonus. Activation is the cheapest of them. Members who never open the offer never cost a point.

Challenges

A challenge asks for a set of actions inside a window and pays on completion. Buy in three categories this month, or visit four times in six weeks. The reward lands only when the member finishes, so the cost follows the behavior. Challenges suit goals that need repeat visits, such as a new habit or a first try of a new category.

Tier pushes

A tier push targets members close to a status line. Some programs also match status from a rival and set a target to keep it. Southwest runs a status match that gives A-List or A-List Preferred status for 120 days. To keep A-List for 12 months, the member flies three round trips in that window or earns 11,500 tier qualifying points from flights (Southwest).

The same terms settle a question every tier push raises. Rapid Rewards bonus points do not count toward tier qualifying points unless Southwest says so. Write that rule down before launch. A bonus that also buys status costs twice, once in points and again in elite benefits.

Win-back and reactivation

A reactivation campaign targets members whose visits are slowing. A win-back campaign targets members already past the lapse line the program sets. The split matters because the offer should grow with the time away. A member who skipped one cycle needs a reminder and a balance update. A member gone for a year needs a reason to return that beats whatever took the program's place.

Time the trigger from each member's own rhythm. A weekly shopper who misses three weeks has broken a habit. A member who travels once a year has not.

Partner promotions

A partner promotion lets members earn or burn with another brand, and the partners share the cost. Delta SkyMiles and Starbucks Rewards members who link their accounts earn 1 mile per eligible $1 at participating Starbucks cafés. They also get Double Stars on days they fly Delta (Delta and Starbucks).

The August 5, 2026 update tightened who earns. Earning miles now needs a qualifying Delta flight in the past 12 months. Miles also stopped posting on Starbucks Card reloads. Eligibility rules like these decide where each partner's funding goes.

How do you plan and target a loyalty campaign?

Start from one behavior and one segment. A campaign that asks everyone for everything cannot be priced or measured. Write a one-page brief before anyone builds a rule. A working brief answers eight questions.

1. Goal: which behavior should change, and by how much, for the campaign to pay back.
2. Audience: who is eligible, written as a rule the system can run.
3. Exclusions: who is left out. Drop members already in a live campaign first. Then drop members under fraud review or with no marketing consent.
4. Window: the start and end dates, with the time zone stated.
5. Offer: the reward, the earn math and the cap per member.
6. Budget: a cost ceiling in points and in money, signed by finance.
7. Measure: the main metric and the size of the holdout.
8. Owner: one named person who can stop the campaign mid-flight.

Target on what members did, then on what they are likely to do next. Recency, visit gaps, distance to the next tier and category history all sit in the program's own data. Members who would buy anyway are the costliest target. Every bonus they collect is margin given away for no change.

Check consent before you size the audience. Starbucks asks members to opt in to marketing email before they can get its personal email offers. To hear about a Double or Triple Star Day, members sign in to the app or opt in to that email (Starbucks). A member the offer cannot reach does not belong in the forecast.

Keep one shared calendar. Two campaigns aimed at the same member in the same weeks spoil both readouts. Nobody can say which one moved the spend.

How should campaign rules and offers be set up?

Rules turn the brief into something a system can run. Each rule should be testable: given this member and this purchase, the outcome is known in advance. Set these before launch.

• Qualifying action: the product, channel, store or spend level that earns.
• Earn math: a multiplier on base points or a flat bonus. A flat bonus fixes the cost per member. A multiplier grows with spend.
• Caps: a limit per member and a total limit for the campaign.
• Activation: whether members opt in first. Opt-in keeps the bonus away from members who never saw the offer.
• Stacking: what happens when two offers apply to one basket.
• Tier credit: whether bonus points count toward status.
• Abuse terms: the right to remove points and suspend accounts.

Write the abuse terms into the offer itself. Starbucks and its partners reserve the right to review all offer activity. They can suspend accounts or remove Stars for suspected abuse or fraud (Starbucks). Those terms are the base for stopping promotion abuse once the campaign is live.

Test every rule against real members before launch. Run last month's purchases through the new rules and count the points each rule would have issued. If the total breaks the budget on old data, it will break it harder live. The campaign exists to raise spend.

How do you control campaign cost and points liability?

Every bonus point is a promise to pay later. Starbucks states the accounting plainly in its annual report. It defers revenue for Stars as each one is earned and records a matching liability in deferred revenue. That deferral is based on the expected value of the reward, net of the Stars it expects will never be redeemed (Starbucks 10-K).

That gives the campaign cost formula. Cost equals bonus points issued, times the share expected to be redeemed, times the cost of one point. Finance should sign the redemption rate and the point cost. The same numbers drive the points liability on the balance sheet.

Four controls keep the cost under the ceiling.

• A cap per member, so one heavy buyer cannot take the budget.
• A total cap with an automatic stop when issued points reach the ceiling.
• Targets above normal spend, so the bonus pays for the extra visit and leaves the usual one alone.
• Partner funding, where the partner pays for the points its offer issues.

How do you measure a loyalty campaign with a holdout?

A holdout is a random slice of eligible members who get no offer. They meet the same rules and live through the same weeks. The gap between the two groups is the campaign's real effect, called incremental lift.

Some messaging tools build this in. Braze, for one, lets a team set a percentage of all users who should not receive any campaigns or Canvases. It then reports uplift against a random treatment sample (Braze). That global holdout measures the whole calendar. A campaign holdout measures one offer. Run both. The global group shows whether the calendar pays, and the campaign group shows whether this offer did.

Three rules keep the readout honest.

• Compare all targeted members with the holdout, even those who ignored the offer. Members who chose to redeem were not a random group. Set them against the rest and the campaign gets credit for a gap it did not cause.
• Measure margin, then subtract the reward cost. Sales lift alone hides offers that cost more than they earn.
• Watch the weeks after the end date. If the treated group buys less than the holdout then, the campaign moved purchases earlier. It did not add new ones.

Size the holdout before launch. It has to be large enough to show the smallest lift that still pays back the reward. The analyst confirms that with a power calculation on past data. Program-wide control groups follow the same logic at a larger scale, as set out in how to measure loyalty program ROI.

A worked example

For illustration, take a grocery program that targets 200,000 members with a challenge. Three visits in 30 days earns 1,000 bonus points. The figures are assumptions. A random 10% holdout, 20,000 members, gets no offer. The other 180,000 get the challenge.

Over the 30 days, the treated group spends $7 more per member than the holdout. At a 30% gross margin, that is $2.10 per targeted member. Across 180,000 members, the extra margin is $378,000.

Of the treated members, 12% finish the challenge. That is 21,600 members and 21.6 million bonus points. If 80% of those points get redeemed at $0.006 each, the reward cost is $103,680. The net gain is $274,320, before message costs.

Now the trap. Suppose finishers spent $60 more than members who did not finish. In this example, most of that gap comes from who chose to finish. Only the holdout comparison gives the true $378,000. A team that read finishers against non-finishers would book a far larger lift and size the next offer wrong.

Which rules and metrics fit each campaign type?

Use this planning table to set up the brief. Each row names the behavior the campaign buys and the rule to settle first. It also names the cost risk to watch and the metric for the readout.

Who runs loyalty campaigns, and what tools do they need?

A campaign touches four desks. The campaign manager owns the brief and the calendar. An analyst builds the audience and reads the holdout. Finance signs the cost ceiling and the redemption rate. The CRM owner ships the messages. Partner promotions add a partner manager, and generous offers add someone from fraud. Each readout then goes to the program's monthly review, part of loyalty program management.

Give one person the stop button. A campaign that runs over budget or draws abuse needs a named owner who can pause it the same day, without waiting for a release.

The tooling follows the same jobs.

• A rules engine the loyalty team edits itself, with a test mode and version history.
• Real-time earn, so members see the bonus at checkout or in the app while the campaign runs.
• Segments drawn from the program's own member data and refreshed before launch.
• A link to the messaging platform, so the offer and the message use one audience list.
• Live monitoring for odd earn patterns during the window.
• Holdout assignment and a readout that subtracts reward cost from margin.

How GRAVTY handles loyalty campaign management

GRAVTY sets campaign rules in the same engine as the base program. Its patented Visual Rules engine lets loyalty managers set earn, burn, tier and bonus logic in a drag-and-drop editor. They simulate it and deploy it without an IT ticket. Every rule is versioned with one-click rollback.

While a campaign is live, AI-Trust flags odd earn and redemption behavior as it happens. Each program can tune the scoring. Agentic AI Compass adds offer testing and anomaly detection. It simulates the outcome before a change goes live. Each recommendation it makes is explainable and auditable.

In ecosystem programs, partners run their own campaigns and settle through integrated invoice management. GRAVTY runs 500M+ members in production. So the loyalty team changes a campaign rule itself and can roll it back in one click.

Frequently asked questions

How often should a loyalty program run bonus points campaigns?

Run them as often as holdout readouts show a net gain, and no more often. A fixed monthly event can shift purchases onto event days. The weeks after each event, read against the holdout, show whether that is happening.

What is the difference between a loyalty campaign and a loyalty program?

The program is the always-on rule set, while a campaign is a time-bound offer on top of it for a chosen group. The program's earn, tier and reward rules set what members expect. Campaigns test changes to that deal before the program adopts them for good.

Should bonus points count toward tier status?

Only when the campaign's goal is status. Southwest's status match terms leave Rapid Rewards bonus points out of tier qualifying points unless they are marked otherwise. Counting bonus points toward status makes one campaign pay twice, in points and in elite benefits.

How large should a campaign holdout group be?

Large enough to detect the smallest lift that would still pay back the reward cost. Work it out before launch with a power calculation on past data. A small lift on a noisy metric needs a bigger holdout than a large lift on a steady one.

How do you stop members from gaming a loyalty campaign?

Cap what one member can earn, and write the right to remove points into the offer terms. Then watch earn patterns while the campaign is live. Abuse tends to show up as clusters, such as many new accounts on one device.

Who should approve a loyalty campaign before launch?

Finance approves the cost ceiling, and the campaign owner approves the rules after a test run on past data. Legal or compliance reviews the terms when the offer adds a new mechanic, such as a sweepstakes or a new partner.