Scheme cost projection and what-if analysis before launch
Most scheme budgets are a rate multiplied by a hoped-for volume. The real cost depends on how many dealers land in each slab, which nobody knows from a rate. The dealers' own purchase history does know. This page shows how to use it before the circular goes out, and how to check afterwards whether the scheme was worth it.
Scheme cost projection means running a draft scheme's rules over the eligible dealers' real purchase history to see what it would have paid, before it is launched. The result is the zero-lift cost: what the scheme pays if dealers buy exactly as they did before. On top of that, a what-if analysis tests how cost changes when dealers near a slab threshold stretch to cross it. After the scheme, effectiveness is measured by comparing payout and volume for the same dealers against the same period a year earlier.
Why a rate times a volume is not a budget
A draft circular says ₹4, ₹6 and ₹8 a bag across three slabs. The budget note multiplies an average of ₹6 by the planned volume. The actual cost depends on the shape of the dealer base: if most volume sits with dealers in the top slab, the average is closer to ₹8, and if the thresholds are just below where many dealers already buy, the scheme pays the higher rate for no change at all. A scheme can only be budgeted by putting each dealer's own numbers through the rules. The budget planning guide covers the wider programme budget; this page is about one scheme at a time.
Step 1: project the draft on real purchase history
Take the eligible dealers and a comparable past period, normally the same month or quarter last year, so that seasonality is like for like. Apply the draft rules to each dealer's actual lifting. The example below is illustrative. The draft is a monthly scheme paid on all bags: nil below 500 bags, ₹4 a bag for 500 to 999, ₹6 for 1,000 to 1,999 and ₹8 for 2,000 and above. There are 500 eligible dealers, and their lifting in the same month last year looked like this.
| Band by last year's lifting | Dealers | Average bags | Total bags | Rate | Projected payout |
|---|---|---|---|---|---|
| Below 500 bags | 220 | 250 | 55,000 | Nil | Nil |
| 500 to 999 | 150 | 700 | 1,05,000 | ₹4 | ₹4,20,000 |
| 1,000 to 1,999 | 100 | 1,400 | 1,40,000 | ₹6 | ₹8,40,000 |
| 2,000 and above | 30 | 3,000 | 90,000 | ₹8 | ₹7,20,000 |
| Total | 500 | 3,90,000 | ₹19,80,000 |
On history alone the scheme costs ₹19.8 lakh. Spread over 3,90,000 bags that is about ₹5.08 a bag, and at ₹350 a bag the volume is worth ₹13.65 crore, so the scheme is about 1.45 percent of value. This is the zero-lift cost: the amount paid if not one dealer changes behaviour. It is the floor of the budget, not the estimate. Three things are visible straight away. The 220 smallest dealers, 44 percent of the base, get nothing and have no reason to engage. The 30 largest dealers take ₹7.2 lakh, over a third of the payout, for volume they already do. And the cost is spread over all bags, although the purpose is only the additional ones.
Step 2: test sensitivity at the slab thresholds
A slab scheme changes behaviour mainly among dealers who are close to a threshold. They are also where the cost jumps. Count the dealers within about 10 percent below each threshold in the history. Suppose 40 of the 150 dealers in the second band lifted between 900 and 999 bags, averaging 950.
If each of them stretches to 1,000 bags, the payout per dealer moves from 950 x ₹4 = ₹3,800 to 1,000 x ₹6 = ₹6,000. That is ₹2,200 more for 50 more bags, or ₹44 for each additional bag. Across 40 dealers the scheme gains 2,000 bags and costs ₹88,000 more, taking the total to ₹20.68 lakh. Whether ₹44 a bag is acceptable depends on your contribution per bag, which only you know. If it is not, move the threshold or change the basis.
The same test works in reverse. If a threshold sits just below where a cluster of dealers already buys, they collect the higher rate with no stretch at all. In the history above, any dealer who already lifted 1,000 to 1,050 bags earns ₹6 a bag for doing nothing new. A threshold placed a little above a cluster asks for a stretch; a threshold placed a little below it is a gift. The slab designer and the incentive slab guide cover threshold placement.
Step 3: compare alternative designs on the same history
Because the history is fixed, two drafts can be compared fairly. Take an incremental version of the same ladder: nil on the first 500 bags, ₹4 on the next 500, ₹6 on the next 1,000 and ₹8 beyond 2,000.
| Band | Dealers | Average bags | Incremental payout per dealer | Band total |
|---|---|---|---|---|
| Below 500 bags | 220 | 250 | Nil | Nil |
| 500 to 999 | 150 | 700 | 200 x ₹4 = ₹800 | ₹1,20,000 |
| 1,000 to 1,999 | 100 | 1,400 | 500 x ₹4 + 400 x ₹6 = ₹4,400 | ₹4,40,000 |
| 2,000 and above | 30 | 3,000 | 500 x ₹4 + 1,000 x ₹6 + 1,000 x ₹8 = ₹16,000 | ₹4,80,000 |
| Total | 500 | ₹10,40,000 |
The incremental design costs ₹10.4 lakh on the same history against ₹19.8 lakh, and the dealer at 950 bags now gains only 50 x ₹4 = ₹200 for reaching 1,000, so the pull at the threshold is much weaker. Neither design is right in general. The projection puts a price on the choice: ₹9.4 lakh a month buys the cliff effect at each threshold. The same method prices a cap, a growth-over-last-year condition or a change of eligibility from all dealers to selected districts. Arithmetic for each scheme type is in how to calculate a dealer scheme payout.
The hypothetical-purchase calculator
Projection answers the brand's question. The what-if calculator answers the dealer's and the sales officer's: if this dealer lifts so many more bags, what does he earn? Take a dealer at 870 bags on the 24th of the month under the retroactive draft. Today he stands to earn 870 x ₹4 = ₹3,480. If he lifts 130 more bags to reach 1,000, he earns 1,000 x ₹6 = ₹6,000. The extra ₹2,520 on 130 bags is about ₹19 a bag, on top of his normal margin. That is a concrete reason to place one more order, and it is a far better conversation than reading out a slab sheet.
The same calculator, run across a territory, tells a sales officer where to spend the last week: the dealers who are close to a threshold and whose past lifting shows they can realistically cover the gap, not the dealers who are furthest behind. The sales officer focus list post describes that ranking, and next-target visibility in the dealer app covers what the dealer sees.
Step 4: measure effectiveness after the scheme
After the period, compare cost and volume for the same dealers against the same period a year earlier. Continuing the example: the 500 dealers lifted 4,20,000 bags in the scheme month against 3,90,000 a year earlier, and the scheme paid ₹23,10,000, which is ₹5.50 a bag.
| Measure | Working | Result |
|---|---|---|
| Volume change, same dealers | 4,20,000 - 3,90,000 bags | 30,000 bags, up 7.7 percent |
| Scheme cost | As paid | ₹23,10,000 |
| Cost per bag lifted | ₹23,10,000 / 4,20,000 | ₹5.50 |
| Cost per additional bag | ₹23,10,000 / 30,000 | ₹77 |
| Growth of dealers outside the scheme | Observed in the same month | 3 percent |
| Bags the eligible dealers would likely have added anyway | 3,90,000 x 3 percent | 11,700 bags |
| Additional bags attributable to the scheme | 30,000 - 11,700 | 18,300 bags |
| Cost per attributable bag | ₹23,10,000 / 18,300 | about ₹126 |
The headline of ₹5.50 a bag and the honest figure of about ₹126 per attributable bag describe the same scheme. The second is the one to compare with contribution per bag. It is uncomfortable, and it is why growth-based and incremental designs exist: they stop paying for the 3,90,000 bags that were coming anyway.
Be clear about the limits. A year-on-year comparison on the same dealers is not a controlled experiment. Price changes, a competitor's supply problem, a late monsoon or an election all move volume. Dealers outside the scheme are a rough control at best, because they were left out for a reason. A scheme can also pull lifting forward, so look at the month after before closing the file. And some dealers split or merge codes between years, so club old and new codes or the base is wrong. The ROI calculation guide discusses attribution in more depth, and the ROI calculator runs the basic arithmetic.
What makes projection hard in practice
- History is not in one place. Last year's invoices are in the ERP, eligibility is in the circular and the clubbing list is in someone's inbox. Projection needs all three together.
- The draft changes five times. Each revision of thresholds or rates means a rerun. If a rerun takes a day in Excel, the fifth version goes out untested.
- Stacking. The new scheme sits on top of live ones. The cost that matters is the combined rate per dealer, not the new scheme alone.
- New dealers have no history. They need an assumption, stated openly, not a silent zero.
In Unotag's dealer scheme engine the cost of a draft is projected on real purchase history before the scheme is saved, a what-if calculator works out the earning for a hypothetical purchase, and the same configuration then runs the live scheme, so the projected and the actual are calculated by identical rules. Invoices come from ERP integration with SAP or Tally, or from Excel upload. The engine does not predict how dealers will respond. That remains a judgement for the commercial team, and the projection gives it a floor and a set of priced scenarios to work from. The dealer scheme engine explainer covers the other parts.
Key takeaways
- Budget a scheme by running the draft over each eligible dealer's real history, not by multiplying a rate by a planned volume.
- The projection on history is the zero-lift cost, the floor. In the example it is ₹19.8 lakh retroactive against ₹10.4 lakh incremental on the same 3,90,000 bags.
- Test the thresholds: 40 dealers stretching from 950 to 1,000 bags cost ₹44 for each additional bag. Compare that with your contribution per bag.
- Afterwards, measure cost per attributable additional bag against the same dealers a year earlier, and be honest that the comparison is not a controlled test.
Frequently asked questions
What is scheme cost projection?
Scheme cost projection is the practice of applying a draft trade scheme's rules to the eligible dealers' actual past purchases to see what the scheme would have paid. It gives the cost if behaviour does not change, which is the minimum budget the scheme needs.
How do you prepare a dealer scheme budget?
Run the draft rules on each eligible dealer's lifting for the same period last year to get the zero-lift cost. Add scenarios for dealers who stretch across slab thresholds, include schemes already live for the same dealers, and apply any cap. The budget is a range, not one figure.
What is what-if analysis for a trade scheme?
What-if analysis changes one input and recalculates: a threshold, a rate, the payout basis or a dealer's purchase quantity. For the brand it prices design choices. For a dealer it shows what a further purchase would earn under the scheme.
How do you calculate the cost of a slab scheme before launch?
Group eligible dealers by the slab their past lifting falls in, multiply each dealer's quantity by the rate that would apply on the stated basis, and add up. Then count the dealers just below each threshold and cost the case where they cross it.
How do you measure the effectiveness of a dealer scheme?
Compare the same dealers' volume in the scheme period with the same period a year earlier, subtract the growth they would probably have had anyway, and divide the scheme cost by the remaining additional volume. Compare that cost per additional unit with your contribution per unit.
What is a good cost per incremental unit for a trade scheme?
There is no universal figure. It has to be below your contribution per unit for the scheme to pay for itself in the period, unless the aim is strategic, such as a launch or defending a territory. State the aim before the scheme, not after.
Can scheme cost projection predict how dealers will respond?
No. It shows what the rules would pay on known purchases and under scenarios you choose. How many dealers actually stretch is a judgement. Past schemes with similar thresholds are the best guide, and the projection keeps that judgement tied to real numbers.