TL;DR Trade promotion optimization (TPO) uses data from past promotions to choose the offer, depth, timing and outlets that will pay best next time. It rests on three numbers: what you would have sold anyway, what the promotion added, and what that extra volume cost. TPO stalls when field data cannot show which outlets actually ran the scheme.
Trade promotion optimization is the practice of measuring each past promotion and using what you learn to plan the next one. It looks back at each scheme, works out how much extra it really sold, and puts the next round of trade spend where that extra volume came cheapest. Every method it uses comes down to three questions. What would we have sold with no promotion? How much more did the promotion sell? And what did each extra case cost, once you count the discount, the free goods, and the sales it took from later weeks or from our other products?
The harder part is knowing what really happened in the market: which outlets were offered the scheme, which ones put up the display, and what the distributor sold to each shop. A model fed with shipment data and guesses about execution will rank your promotions with great confidence and still get them wrong.
Trade promotion optimization vs trade promotion management
Trade promotion management (TPM) runs promotions. It covers the budget, the calendar, setting up each scheme, telling the trade about it, and checking and paying claims. Trade promotion optimization decides which promotions are worth running. It forecasts what a planned promotion will add and, once the promotion ends, measures what it really did.
| Trade promotion management (TPM) | Trade promotion optimization (TPO) | |
|---|---|---|
| Question it answers | Is the promotion set up, run and paid correctly? | Which promotion should we run, where, and how deep? |
| Main users | Trade marketing, sales operations, finance | Trade marketing, revenue management, analysts |
| When it works | Before, during and after each promotion | When planning, and after each promotion ends |
| What goes in | Budgets, scheme rules, outlet lists, claims | Sales history, promotion records, execution proof, costs |
| What comes out | Live schemes, accruals, settled claims | Baselines, uplift forecasts, return by promotion, the next plan |
Most brands need TPM first. TPO learns from records of past promotions, and those records only exist when promotions are set up, run and claimed in one place. If your schemes still live in emails and spreadsheets, start there. Our guide to trade promotion covers how to plan and run promotions. This page is about measuring them and making the next one better.
What data does trade promotion optimization need?
A TPO model is only as good as the records behind it. Before you trust any forecast, check that you hold these five kinds of data, ideally by product, by outlet or account, and by week.
- Sales history. Weekly sales outside promotion periods, covering every season at least once, so the model can learn your normal pattern.
- Secondary sales by outlet. What the distributor sold to each shop, not only what you shipped to the distributor. Primary shipments show who stocked up. Secondary sales show what moved toward the shopper.
- Promotion records. The mechanic, the depth, the start and end dates, the products, the outlets targeted and who paid for it.
- Execution proof. Whether the rep offered the scheme at the order, whether the display went up, and whether the outlet claimed. Without it you cannot tell a weak offer from an offer nobody heard about.
- Costs. Discounts, free goods, display fees and claims paid, each matched to the promotion that caused it.
Chains in modern trade may share till data, and you can buy syndicated data in many markets. Use it where you have it. In general trade, the distributor’s invoices to each shop are usually the closest thing to sell-out, because shipments only show how much stock the trade took. That is why a distributor management system that records secondary sales is often the first data project in a TPO program.
How does trade promotion optimization work? Five methods
These methods build on each other. You need a sound baseline before any of the others mean much.
1. Baseline and uplift
The baseline is what you would have sold with no promotion, given the season, the outlets you cover and the trend. Uplift is the gap between actual sales during the promotion and that baseline. Get the baseline wrong and every later number is wrong too. Set it too low and every promotion looks like a winner.
2. Post-event analysis
After each promotion ends, compare the forecast with what happened. Then split the volume into its sources: new demand, sales taken from your other products, sales pulled forward from later weeks, and normal sales you would have made anyway. Put the result into a simple profit and loss for that promotion. Skip this step and the next plan repeats the last one’s mistakes.
3. Price elasticity
Elasticity tells you how much volume moves when the price moves. It answers the depth question: does a deeper discount bring enough extra volume to pay for itself? Past a certain depth, the extra discount mostly goes to buyers who would have bought anyway.
4. Product switching and pull-forward
Two effects make promotions look better than they are. First, shoppers and shopkeepers switch from your other products to the promoted one, so the brand sells no more overall. Second, retailers stock up during the deal and then order less for weeks afterwards. Distributors do the same with schemes on primary orders, which is called forward buying, and shoppers do it at home, which is called pantry loading. Measure the weeks after a promotion as well as the weeks during it, and the dip shows up.
5. Scenario planning
Once you have baselines and past uplift, scenario planning lets you test a promotional plan before you spend on it. Change the mechanic, the depth, the timing or the outlet list, and compare the forecast extra volume, cost and profit. The aim is to drop the options that clearly lose money before they reach the calendar.
An example of how the methods change a decision
Take a made-up beverage brand that runs a buy-ten-get-one-free scheme for a month in a group of grocery shops. Volume that month is well above normal, so the scheme looks like a success. Post-event analysis tells a different story. A large share of the extra volume came out of the brand’s own bigger pack, and shops ordered less than normal the month after. Once those two effects and the cost of the free goods are counted, the scheme barely broke even. A smaller display payment in similar shops caused no switching and no dip, and it paid back better for each extra case it sold.
Trade promotion analysis: which measures show a promotion paid?
Judge each promotion on the extra volume it added over the baseline. Read these measures together. Any one alone can make a losing promotion look good.
| Measure | What it tells you | How to work it out |
|---|---|---|
| Incremental volume | What the promotion really added | Sales during the promotion minus the baseline, minus the dip in the weeks after |
| Promotion ROI | Whether the trade investment made money | Subtract the full promotion cost from the gross profit on the incremental volume, then divide by the full promotion cost. Above zero, the promotion made money |
| Cost per incremental case | How cheaply the promotion bought extra volume | Full promotion cost divided by incremental cases |
| Post-promotion dip | How much volume was borrowed from later weeks | Baseline minus actual sales in the weeks after the promotion |
| Execution rate | Whether the trade actually ran it | Outlets where the scheme was offered or the display was verified, divided by outlets targeted |
| Claim accuracy | Whether the money went where it should | Claims matched to verified sales or proof, divided by claims paid |
Use cost per incremental case to compare mechanics on the same product. Use promotion ROI to compare across products and categories, because a cheap case of a low-margin product can still lose money. Read both next to the execution rate. A promotion with poor returns and poor execution needs fixing in the field before you decide to drop it.
Agree who owns each number
Sales, finance and trade marketing often judge the same promotion by different numbers. Sales looks at volume, finance looks at cost, and trade marketing looks at reach. In larger brands, trade promotion optimization sits inside revenue growth management, next to pricing and pack decisions, which adds a fourth view. Settle it before the first analysis. Give one team the baseline, usually revenue management or trade marketing. Let finance own the cost side, and let sales own the execution rate. Then everyone argues from the same promotion performance report.
Why optimization fails without clean field data
A TPO model assumes that a promotion planned for a set of outlets ran in those outlets, unless execution data says otherwise. In general trade that assumption often fails. The rep forgot to mention the scheme, the distributor applied it at the invoice instead of at the order, the display stayed in the back room, or the shop never claimed. The model then records a weak response and tells you to cut a promotion that was never properly tried.
Three habits fix most of this:
- Show the scheme at the order. When the rep’s sales force automation app shows the live schemes for that outlet and applies them as the order is taken, the shopkeeper hears the offer while deciding how much to buy. The record also shows that the scheme was offered.
- Get proof of display. A geo-tagged shelf photo taken during a retail audit shows whether the display went up. Visual merchandising tools can score those photos so nobody has to check them by hand.
- Check claims against secondary sales. Match each claim to the distributor’s invoices to that outlet before you pay it.
AAVA Brands, a fast-growing FMCG company in Nigeria that sells through general and modern trade, had no single view of primary and secondary sales before it brought its field sales and distributor data onto BeatRoute. It went on to grow store-level sellouts by 25–30% (read the AAVA Brands case study). That result came from better execution, and the same outlet data is what a TPO model needs.
Reach matters too, because a scheme cannot work in an outlet your team never talks to. Valvoline Cummins lifted what it calls “customer connect” from 25–30% to 70%+ with BeatRoute (read the Valvoline Cummins case study). For the field side of running schemes, see how to run trade promotions in retail distribution.
What AI and predictive models add to trade promotion optimization
AI is useful in TPO where the work is too big for a person with a spreadsheet. A predictive model can:
- build baselines for every product and outlet cluster at once
- find which features of past promotions went with higher uplift
- run far more scenarios than a team could test by hand
- flag outlets where claims run well above or below sales while the promotion is still live
AI also has clear limits. It learns from history, so it repeats whatever the records got wrong. If the data does not show whether a scheme was offered in an outlet, the model treats a scheme nobody heard about as a scheme nobody wanted. It also has little to say about a mechanic you have never run. Treat its forecast as a starting point that a person who knows the trade signs off.
How trade promotion optimization changes by channel: modern trade vs general trade
Modern trade. You deal with a small number of chains. Promotions are agreed account by account, and many chains share till data. TPO here works at the level of account, product and week. Retailers call their side of this work retail promotion optimization, and many run their own models. The main levers are depth, timing against the chain’s own calendar, and display fees. In-store promoters and display checks decide whether the agreed promotion is actually visible on the shelf.
General trade. Here the brand reaches thousands of small shops through distributors, usually with no scanner data. You cannot model each shop on its own, so group outlets into clusters by type, size and area. Then measure each cluster against similar outlets that did not get the scheme. The distributor’s secondary sales stand in for sell-out. The levers differ too: quantity slabs, free goods, and period purchase schemes that reward what a shop buys across a month or a season.
Many brands sell through both channels. Keep the two analyses apart. A promotion that pays in a chain can lose money in small shops, and the other way round.
How trade promotion optimization differs by industry
- FMCG and food and beverage. Many products and frequent schemes, so switching between your own products and pull-forward matter most. Watch pack sizes and flavours that compete with each other.
- Pharmaceuticals. Schemes run with pharmacies and wholesalers, and rules often restrict promoting medicines directly to consumers. Measure by secondary sales to each pharmacy, and check that every scheme stays within those rules.
- Building materials, paints and lubricants. Dealer slab schemes and period schemes, often with loyalty schemes for painters, mechanics or contractors. Buying cycles are longer, so the dip after a promotion takes longer to show. Measure over a longer window.
The tools a trade promotion optimization program needs
No single tool covers all of TPO for every brand. It helps to think in kinds of tool, by what each one does and who it suits.
- Planning and optimization suites. They build baselines, forecast uplift and run scenarios, and many also manage the promotion calendar, budgets and accruals. They suit large brands with long modern trade calendars and an analytics team to run them.
- Field execution and distributor systems. These put schemes in the rep’s order app and the distributor’s billing, capture display proof and record secondary sales by outlet. They suit brands that sell through distributors to many small shops.
- Spreadsheets and dashboards. Fine for a first post-event analysis of a few promotions. They break down once promotions, outlets and products multiply and nobody owns the baseline.
Whatever kind you look at, ask the vendor:
- How do you build the baseline, and can my team see and adjust it?
- Do you measure the dip after a promotion and switching between my own products?
- At what level do you model: outlet, cluster, account or region?
- Where does execution proof come from, and does it feed the analysis?
- Can you work from secondary sales when I have no sell-out data?
- How soon can I see a first post-event analysis on my own data?
Where BeatRoute fits. BeatRoute is the execution and secondary sales side of TPO. Its field sales app shows reps the live schemes for each outlet and applies them at the order. Its distributor management system holds the scheme and claim rules and records what each distributor sells to each shop, and its retail audit tools record whether displays went up. Goal-Driven AI turns a scheme goal, such as must-sell compliance in a set of outlets, into targets and nudges for each rep. That gives a TPO model clean data on what ran, where, and what it sold. If your main need is planning a large modern trade promotion calendar with detailed price models, a dedicated planning suite is the better fit, and BeatRoute’s outlet data can sit alongside it through integrations.
Run a pilot before you scale
You can test TPO on one promotion cycle before you change the whole budget.
- Pick one category and one region where you run promotions often.
- Choose two or three mechanics to compare, such as a price-off, free goods and a display payment.
- Split similar outlets into test groups and a control group that gets no scheme.
- Agree the success measure and the bar it must clear before you start, for example a cost per incremental case below your current average.
- Run one full cycle with execution tracked: scheme offered at the order, display photos taken, claims checked against secondary sales.
- Read the result against the control group, including the weeks after the promotion ends.
Before you scale, ask four questions. Did the scheme actually run in the test outlets? Did the winner still beat the control after the dip? Did it hold across outlet clusters? Can your team repeat the analysis without outside help?
If execution data is your weak link, book a demo to see how BeatRoute puts schemes in the rep’s hand and tracks secondary sales by outlet.
Frequently Asked Questions
What is trade promotion optimization (TPO)?
Trade promotion optimization uses data from past promotions to choose the offer, depth, timing and outlets that will pay best. It works out what each promotion added over normal sales and what that extra volume cost, then moves budget toward the promotions that bought extra volume most cheaply.
What is the difference between TPM and TPO?
Trade promotion management (TPM) runs promotions: budgets, calendars, scheme setup and claims. Trade promotion optimization (TPO) decides which promotions to run by forecasting and measuring what each one adds. TPO depends on the clean records that TPM produces.
What data do you need for trade promotion optimization?
You need weekly sales history, secondary sales by outlet, a record of each promotion’s mechanic, depth, dates and outlets, proof that it ran, and its full cost. Sell-out or till data helps where you can get it.
How do you measure trade promotion effectiveness?
Measure incremental volume against a baseline and subtract the dip after the promotion. Then work out promotion ROI and cost per incremental case. Read both next to the execution rate, so you do not blame the offer for a scheme that never reached the outlet.
Can TPO work in general trade without scanner data?
Yes. Use the distributor’s secondary sales to each outlet as the sales measure, group similar outlets into clusters, and compare outlets that got the scheme with similar ones that did not.
What role does AI play in trade promotion optimization?
AI speeds up baseline forecasts, finds patterns across many past promotions, runs scenarios, and flags outlets where a scheme is under-claimed or over-claimed. It cannot fix missing data. If the records do not say whether a scheme ran, the model learns the wrong lesson.
How long should a trade promotion run?
Long enough to cover at least one full ordering cycle for the outlets in the scheme, so every shop has a chance to buy on it. Short enough that the deal does not become the price the trade expects. Post-event analysis of past promotions of different lengths is the best guide for your category.
