TL;DR Trade promotion optimization (TPO) is how a brand learns which of its trade schemes made money and which only gave margin away. You compare each scheme with what the same outlets would have bought anyway, then count the full cost of the extra cartons. In Africa the weak link is rarely the maths. It is knowing which dukas, spazas and kiosks actually got the offer.
Your distributors ran a scheme last month and orders went up. Was it a good scheme? Many sales heads in Lagos, Nairobi or Accra cannot answer that with confidence. TPO gives you a way to answer. You test each scheme against normal sales, find out what the extra volume cost, and then back the schemes that were cheapest per extra carton.
Every TPO method rests on three questions:
- What would these outlets have bought with no scheme?
- How much more did they buy because of it?
- What did each extra carton cost? Count the discount, the free stock, and orders pulled from later weeks or from your other products.
The answers depend on facts from the street. Which shops were told about the scheme? Who put up the display? What did the distributor invoice to each stockist and retailer? If you feed a model shipment figures and guesses, it will rank your schemes with total confidence and still be wrong.
TPM and TPO: what is the difference?
Trade promotion management (TPM) is the running of a scheme. It handles the budget, the calendar, the rules, the message to distributors and stockists, and the claims. TPO is the judging of a scheme. Before launch it forecasts what a scheme should add. After the scheme ends it shows what it did add.
| Trade promotion management (TPM) | Trade promotion optimization (TPO) | |
|---|---|---|
| Question it answers | Was the scheme set up, run and paid the right way? | Which scheme should we back, in which outlets, and at what depth? |
| Main users | Trade marketing, sales operations, finance | Trade marketing, commercial heads, analysts |
| When it works | Before, during and after each scheme | At planning time, and again once the scheme closes |
| What goes in | Budgets, scheme rules, outlet lists, claims | Past sales, scheme records, proof of execution, costs |
| What comes out | Live schemes, accruals, settled claims | Baselines, forecast uplift, return per scheme, next plan |
Start with TPM. TPO needs a trail of past schemes, and that trail only exists if schemes are set up, run and claimed in one place. If yours are scattered across WhatsApp groups, emails and Excel sheets, fix that first. Our trade promotion guide shows how to plan and run schemes. This page is about judging them and doing better next time.
Which data does TPO need?
A forecast is only as good as its inputs. Before you trust one, check that you can pull these five things, split by product and outlet, ideally by week.
- Sales history. Normal weeks, with no scheme running, across every season. Think of the festive rush, the rains and the planting or harvest cycle in your region. The model needs to see your usual rhythm.
- Secondary sales by outlet. What your distributor sold on to each shop. Shipments to the distributor only show who loaded stock. Secondary sales show what went out toward the shopper.
- Scheme records. The type of offer, how deep, the dates, the products, the target outlets and who funded it.
- Proof of execution. Did the rep make the offer at the order? Did the display go up? Did the outlet claim? Without this, you cannot tell a poor offer from one that nobody heard.
- Costs. Discounts, free goods, display fees and claims paid, each tied to the scheme that triggered it.
Supermarket chains may share till data. Use it when you can. The duka, the kiosk and the provision store have no scanner at all. For them, the distributor’s invoice to the shop is the nearest thing to a sell-out figure. That is why a distributor management system that captures secondary sales is so often the first step of a TPO program.
Five TPO methods, step by step
Each method sits on the one before it. Without a fair baseline, the rest mean little.
1. Baseline and uplift
The baseline is your estimate of normal sales for the same outlets, season and trend. Uplift is actual sales minus baseline. A baseline set too low makes every scheme look brilliant, and every number after it carries the error.
2. Post-event analysis
Once a scheme closes, set the forecast beside the result. Then break the extra volume into parts. These are truly new demand, sales borrowed from your other products, orders pulled forward from coming weeks, and sales that would have happened anyway. Turn this into a short profit and loss for the scheme. Teams that skip it end up repeating last quarter’s mistakes.
3. Price elasticity
Elasticity measures how far volume moves when price moves. It tells you whether a deeper cut earns back its cost. Beyond some point, the extra discount only rewards buyers who were coming anyway. Where the naira, shilling or cedi is under pressure, retailers watch every price change closely, so depth matters even more.
4. Product switching and pull-forward
Two things flatter a scheme. One is switching: a shopkeeper picks your promoted pack instead of your other pack, so the brand gains nothing overall. The other is stocking up. A wholesaler or stockist loads up during the deal and then buys little for weeks. Trade people call this forward buying. Measure the weeks after the scheme, and the dip appears.
5. Scenario planning
With baselines and past uplift in hand, you can try a plan on paper first. Change the offer, the depth, the dates or the outlet list and compare forecast volume, cost and profit. Cut the options that clearly lose money before they reach the calendar.
A worked example
Picture a beverage brand that offers buy ten, get one free for a month through stockists who supply provision stores and kiosks. Orders jump, and everyone calls it a win. The post-event view is less kind. Much of the jump came from the brand’s own larger pack, and shops bought less the next month. After you subtract those two effects and the cost of the free stock, the scheme only just broke even. A small display payment in similar shops caused no switching and no dip, so each extra carton cost less.
Which numbers show that a scheme paid?
Judge a scheme by the volume it added above baseline. Use the measures below together, because any single one can make a losing scheme look fine.
| Measure | What it shows | How to calculate it |
|---|---|---|
| Incremental volume | What the scheme truly added | Sales in the scheme period, less baseline, less the dip afterwards |
| Promotion ROI | Whether the spend made money | Gross profit on incremental volume, minus full scheme cost, divided by full scheme cost. Above zero means it paid |
| Cost per incremental carton | How cheaply you bought extra volume | Full scheme cost divided by incremental cartons |
| Post-promotion dip | Volume borrowed from later weeks | Baseline minus actual sales in the weeks after |
| Execution rate | Whether the trade really ran it | Outlets with the offer made or the display verified, divided by outlets targeted |
| Claim accuracy | Whether money went to the right place | Claims backed by verified sales or proof, divided by claims paid |
Compare mechanics on one product with cost per incremental carton. Compare across products with ROI, since a cheap carton of a thin-margin product can still lose money. Always read both beside the execution rate. If returns and execution are both weak, repair the field work before you kill the scheme.
Decide who owns each number
Sales watches volume, finance watches cost and trade marketing watches reach. Each team may judge the same scheme differently. Agree the owners before the first review. Let trade marketing or the commercial team own the baseline, finance own cost and sales own the execution rate. Then everyone reads one report.
Why TPO breaks without clean field data
A model assumes that a scheme planned for certain outlets ran in them. In general trade, that is often untrue. The rep forgot to mention it. The distributor applied it at invoicing, not at the order. The display never left the storeroom. The shop never claimed. The model sees weak sales and tells you to drop a scheme that was never really tried.
Open markets add a second problem. Sub-wholesalers at hubs such as Onitsha and Gikomba restock daily and carry discounted stock across territory lines. Your uplift then appears in the wrong place. Outlet-level records show where stock went, so you reward the outlets that earned the volume and not the ones that only resold it.
Three habits fix most of it:
- Put the scheme in front of the shopkeeper at the order. When the rep’s sales force automation app lists the live schemes for that outlet and applies them as the order is typed, the shopkeeper hears the offer while buying. The record proves it too. When prices move, the app also stops reps from quoting an old discount.
- Get a photo of the display. A geo-tagged shelf picture from a retail audit shows whether the display is up. A time-stamped visit also ends the ghost visit, where a call is marked done from the car. Visual merchandising tools can score the photos, so no manager has to check each one.
- Test every claim against secondary sales. Match the claim to the distributor’s invoices for that outlet before payment. Digital claims with proof attached also clear faster, which matters because the distributor often funds the scheme until you pay.
AAVA Brands is a fast-growing FMCG company in Nigeria that sells through general and modern trade. Before BeatRoute, it had no single view of primary and secondary sales. After it moved field sales and distributor data onto the platform, it grew store-level sellouts by 25–30% (AAVA Brands case study). Better execution drove that result, and a TPO model needs the same outlet data. For the field side of schemes, read how to run trade promotions in retail distribution.
Where AI helps TPO, and where it does not
AI earns its place when the job is too big for one analyst and a spreadsheet. A predictive model can:
- set baselines for every product and outlet cluster in one pass
- spot which traits of past schemes went with higher uplift
- test far more scenarios than a team could try by hand
- flag outlets whose claims run far above or below their sales while the scheme is still open
It has limits. It learns from history, so it copies the gaps in your records. If nothing shows whether an outlet was offered the scheme, the model reads “nobody heard” as “nobody wanted”. It also cannot say much about an offer you have never run. Take its forecast as a draft, and let someone who knows the trade approve it.
TPO in modern trade and in general trade
Modern trade. A few supermarket chains take most of the effort. Schemes are agreed account by account and many chains share till data, so TPO works by account, product and week. The levers are depth, timing against the chain’s own calendar, and display fees. In-store promoters and display checks decide whether the deal is visible on the shelf.
General trade. You reach thousands of open-market stalls, dukas, spazas, kiosks and provision stores through distributors and wholesalers, with no scanner data. You cannot model each one, so group them into clusters by type, size and area. A small share of outlets often drives most of the volume, which shows where a scheme can pay. Compare each cluster with similar outlets that got no scheme, and use distributor invoices as the stand-in for sell-out. The levers differ as well: quantity slabs, free goods and period purchase schemes that reward a shop for buying steadily across a month or season.
If you sell in both channels, keep the two reviews apart. A scheme that pays in a supermarket can lose money in small shops, and the reverse is also true.
TPO by industry
FMCG and food and beverage. Many products and frequent schemes, so switching and pull-forward do the most damage. Watch pack sizes and flavours that compete with each other.
Pharmaceuticals. Schemes go through pharmacies and wholesalers, and rules often limit promoting medicines to the public. Measure through secondary sales to each pharmacy and keep every scheme inside those rules.
Building materials, paints and lubricants. Dealer slab schemes and period schemes, often with rewards for painters, mechanics or contractors. Buying cycles are longer, so the post-scheme dip shows up later. Use a longer measuring window.
Which tools does a TPO program need?
No one tool covers all of TPO for every brand. Think in kinds of tool, by what each does and who it suits.
- Planning and optimization suites. They build baselines, forecast uplift and run scenarios, and many also handle calendars, budgets and accruals. They suit large brands with long modern trade calendars and an analytics team.
- Field execution and distributor systems. They carry the scheme into the rep’s order app and the distributor’s billing, collect display proof and record secondary sales by outlet. They suit brands that reach many small shops through distributors.
- Spreadsheets and dashboards. Good enough for a first look at a few schemes. They fail once schemes, outlets and products multiply and nobody owns the baseline.
Put these questions to any vendor:
- How is the baseline built, and can my team see it and adjust it?
- Do you measure the dip after a scheme and switching between my own products?
- At what level do you model: outlet, cluster, account or region?
- Where does proof of execution come from, and does it feed the analysis?
- Can you work from secondary sales when I have no sell-out data?
- Does the app work on a basic phone with weak signal and sync when the rep is back online?
- How soon can I see a first post-event result on my own data?
Where BeatRoute fits. BeatRoute covers execution and secondary sales. Its field app shows each rep the live schemes for the outlet and applies them at the order. Its distributor management system holds scheme and claim rules and records what each distributor sells to each shop. Its retail audit tools record whether displays went up. Goal-Driven AI turns a scheme goal, such as must-sell lines in a set of outlets, into daily targets and nudges for each rep. It works with the distributors and reps you already have and does not replace them. A TPO model then gets clean facts on what ran, where, and what it sold. If you mainly need to plan a big supermarket calendar with detailed price models, a dedicated planning suite fits better. BeatRoute outlet data can sit beside it through integrations.
Try TPO on one scheme cycle first
You do not need to change the whole budget. Test it on a single cycle.
- Choose one category and one region where you run schemes often.
- Pick two or three offers to compare, for example a price-off, free goods and a display payment.
- Split similar outlets into test groups and one control group that gets no scheme.
- Fix the success bar before you start, such as a cost per incremental carton below your current average.
- Run one full cycle and track execution: offer made at the order, display photos taken, claims checked against secondary sales.
- Read results against the control group, including the weeks after the scheme ends.
Before you scale, ask four things. Did the scheme really 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 work alone?
Is proof of execution your weak spot? Book a demo and we will show how BeatRoute puts schemes in the rep’s hand and tracks secondary sales by outlet.
Frequently asked questions
What is trade promotion optimization (TPO)?
TPO is a way of grading your past trade schemes. For each one you find out how much extra it sold over normal sales and what that extra cost. You then shift the next budget to the schemes that gave the cheapest extra cartons.
What is the difference between TPM and TPO?
TPM is the day-to-day work of running schemes: budgets, calendars, rules and claims. TPO is the review that tells you which schemes deserve a repeat. TPO only works well on top of tidy TPM records.
What data do you need for trade promotion optimization?
Five things. Weekly sales in normal periods and secondary sales by outlet. A record of every scheme (offer, depth, dates, outlets). Proof that it ran, and its full cost. Till data from supermarkets is a bonus.
How do you measure trade promotion effectiveness?
Take the extra volume above baseline and subtract the dip in later weeks. Then work out ROI and cost per incremental carton. Check both against the execution rate, so a scheme that never reached the shops is not blamed on the offer.
Can TPO work in general trade without scanner data?
Yes. Use the distributor’s invoices to each shop as your sales figure. Group similar dukas, spazas, kiosks and provision stores into clusters, then compare those that got the scheme with those that did not.
Why do trade promotions leak money in African trade?
Once stock leaves the depot, most brands lose sight of it. They pay for a display or a discount and cannot confirm it reached the shop. Weak scheme design is rarely the main cause. Missing proof is.
What role does AI play in trade promotion optimization?
AI builds baselines faster, finds patterns across many old schemes, tests scenarios and flags outlets whose claims look off. It cannot invent missing facts. If your records do not show whether a scheme ran, it will draw the wrong lesson.
How long should a trade promotion run?
Cover at least one full ordering cycle so every shop gets a chance to buy. Stop before the deal turns into the price the trade expects. Results from your own past schemes of different lengths are the best guide.
Is BeatRoute a CRM?
No. BeatRoute is a field sales (SFA) and distributor management (DMS) platform. SFA is the rep’s app in the field. DMS keeps distributor stock, orders and claims in view.
