Promotions produce some of the most immediately visible numbers on a sales report. Cut the price by 20%, Orders increase, stock moves faster and the dashboard quickly shows a sales spike.
But those numbers do not yet answer the most important question:

Would some of those customers have purchased anyway?

If most of them were already planning to buy, the promotion may have created relatively little additional demand. The business may simply have collected less Revenue or Margin from transactions that were likely to happen anyway.
Promotion measurement should therefore go beyond “How much did we sell during the promotion?”

The better question is:| “How much additional Sales and Profit did the promotion actually create?”

Did the promotion really generate more sales?

Short answer: measure incremental sales, not total sales during the promotional period.

Consider a simple example:
Expected normal sales = 1,000 units
Sales during Promotion = 1,300 units

It would be an overstatement to say the promotion generated 1,300 sales because some of those 1,000 baseline units were likely to happen anyway.
At a simple level:
‍Incremental Sales = Actual Sales during Promotion − Baseline Sales without Promotion

In this example:
1,300 − 1,000 = 300 Incremental Units

The critical concept is the Baseline.
Baseline Sales represent the sales expected to occur without the promotion, a situation that cannot be directly observed once the promotion actually runs. In causal analysis, this unobserved alternative is the Counterfactual. In the real world, the same store cannot simultaneously run the same week both with and without a promotion. Promotion Measurement is therefore largely about constructing the most credible comparison possible.

Why can a simple Before vs. During comparison be misleading?

The easiest analysis is: Sales before Promotion vs. Sales during Promotion
This can provide a useful signal, but it does not establish that the promotion caused the entire change.
Other factors may have changed between the periods:

• Payday
• Holidays
• Seasonality
• Competitor Promotions
• Advertising Spend
• Distribution
• Stock Availability
• Product Launches
• Weather
• Pricing or Product Mix

If a promotion runs during a naturally strong end-of-month period, comparing it with a weaker period earlier in the month can make the promotion appear more effective than it really was.
So: Before–After analysis shows that Sales changed.

It does not by itself prove that: The promotion caused the full change.

How can you estimate Baseline Sales?

There is no single method that fits every business. The right approach depends on available data and how important the decision is.

Method 1: Historical Baseline

A smaller business can start with normal sales from comparable periods, such as:

• Average Sales over the previous 4–8 weeks
• The same day of the week
• The same period last year
• A seasonally adjusted benchmark where necessary

This is simple and practical, but the market or business may have changed between periods.

Method 2: Compare with a non-promoted group or location

If the business has multiple stores, locations or reasonably comparable customer groups, one group may provide a reference.
For example:

Store A runs the Promotion
Store B, with a similar historical pattern, does not

Comparing changes across the two stores may account for some external factors better than a simple Before–After comparison. The key limitation is whether the comparison groups were genuinely similar.

Method 3: Controlled Experiment

For higher-value decisions, a properly designed test that separates a Treatment Group receiving the promotion from an appropriate Control Group can provide stronger evidence of Incrementality. The same basic principle is used in marketing Lift Experiments, where outcomes from an exposed group are compared with a group that did not receive the intervention. The stronger the claim that a promotion caused an outcome, the stronger the measurement design should be.

Do not stop at Incremental Sales - measure Incremental Profit

A promotion can sell more units and still reduce profit.
Suppose:
Regular price = THB 100
Cost per unit = THB 60
Gross Profit per unit = THB 40

During the promotion, the price is reduced to THB 80.
Gross Profit per unit becomes:
80 − 60 = THB 20

If normal sales are 1,000 units:
Gross Profit = THB 40,000

During the promotion, sales rise to 1,300 units:
Gross Profit = THB 26,000

Unit Sales increased 30%, yet Gross Profit in this example declined.
That does not automatically make the promotion a bad decision. The objective might be Customer Acquisition, Trial or Inventory Clearance. But if the objective is near-term profitability, Sales Lift alone is not enough.

Who used the promotion?

Even when Orders increase, it matters which customers generated them.

• New Customers - If the promotion attracts first-time buyers, examine whether they return at normal prices later. A large number of New Customers does not automatically make a promotion profitable if the discount is expensive and subsequent Repeat Purchase is weak.
• Existing Customers - if most discounts go to customers who regularly purchase anyway, the business may experience Discount Leakageม giving up Margin on transactions that had a relatively high probability of occurring without the discount. Customer type alone cannot prove that the purchase would have happened anyway. A Baseline or stronger comparison is still required.
• Dormant Customers - Customers who have not purchased for some time can be particularly informative. If a promotion reactivates them and their purchasing continues afterwards, the business implication may be different from simply discounting purchases for already-active customers.

Watch for Cannibalization: The promoted SKU may grow while the business barely does

Suppose a retailer sells products A and B, which customers consider substitutes.
Before the promotion:

A = 500 units
B = 500 units
Total Category = 1,000

During a promotion on A:
A = 800 units
B = 300 units
Total Category = 1,100

If you look only at product A, Sales increased by 300 units.
At the Category level, however, total volume increased by only 100 units.
Some of A's growth may have come from customers switching away from B.

That is Cannibalization.

Promotion effectiveness should therefore be examined at the right level:
- SKU
- Brand
- Category
- Channel or
- Total Business

Growth at one level can simply represent movement from another part of the portfolio.

What happens after the promotion matters too

A promotion may cause customers to bring a future purchase forward. Imagine customers normally buy a product once a month. During a Buy 2 Get 1 Free promotion, they stock up. Sales spike during the promotion. The following month, sales fall because customers already have enough product. This is one possible form of Purchase Timing Shift or Stockpiling.

For products that can be stored, it is useful to examine: Pre-Promotion → Promotion → Post-Promotion rather than analysing only the promotional window. If the promotional spike is followed by a significant post-promotion decline, some of the apparent Lift may have been shifted from future demand rather than entirely new demand.

Which metrics should a Promotion Review include?

A practical review can start with seven measures:

1. Actual Promotion Sales
How much was sold during the promotion?

2. Baseline Sales
How much would likely have been sold without the promotion?

3. Incremental Sales / Units
How much volume was generated above Baseline?

4. Incremental Revenue
How much additional Revenue came from the incremental volume?

5. Gross Margin / Incremental Profit
What remains after the discount and promotion costs?

6. Customer Mix
How much came from New, Existing or Dormant Customers?

7. Post-Promotion Sales
Did Sales return to normal, decline or remain elevated after the campaign?
Not every promotion requires all seven. Choose the measures based on the objective.

Define the objective before deciding whether a promotion “worked”

The same promotion can be successful under one objective and unsuccessful under another.

• If the objective is Incremental Profit - Measure Incremental Gross Profit or Contribution after Discounts and Promotion Costs.
• If the objective is Customer Acquisition - Examine Incremental New Customers, Acquisition Cost and subsequent customer behaviour.
• If the objective is Trial - Measure how many customers actually tried the product and what happened afterwards.
• If the objective is Inventory Clearance - Lower Margin may be acceptable if the business achieves the required inventory reduction and carrying-cost benefit.
• If the objective is Traffic - Higher store or website traffic may matter, but examine Conversion and Basket Economics afterwards. There is therefore no single metric that defines Promotion Effectiveness for every objective.

Promotion ROI should use incremental value, not the entire promotional Revenue

A common mistake is to take all Revenue generated during a promotion and compare it with Promotion Cost. The problem is that some of that Revenue may have occurred without the promotion. When the objective is to understand additional business impact, Incremental Revenue or Incremental Profit is usually the more relevant basis.

Conceptually: Promotion ROI = Incremental Profit attributable to Promotion ÷ Promotion Investment
‍The exact calculation depends on which costs are included, such as Discount Cost, Media Spend, Trade Fees, Coupon Costs and Operating Costs. ROI comparisons become misleading when different campaigns use different definitions of cost or incremental value.

What if you cannot run an experiment yet?

You do not need a sophisticated analytics system to improve Promotion Measurement.
Start with a simple Promotion Sheet containing:
- Promotion Period
- Product / SKU
- Regular Price
- Promotion Price
- Baseline Sales
- Actual Sales
- Incremental Units
- Gross Margin
- New vs Existing Customers
- Post-Promotion Sales
- External Factors

Use consistent definitions from one campaign to the next. Over time, the business can begin comparing which promotions:
- Generate Incremental Volume
- Mostly shift purchase timing
- Reduce Margin too heavily
- Attract new customers or
- Produce behaviour that continues after the discount ends

This creates far more useful institutional knowledge than storing only “Sales during Campaign.”

What promotion data can and cannot tell you

Before–During–After Sales data can reveal patterns and support a basic Baseline. But without a suitable Control or comparison, it is usually too strong to claim that the promotion caused the entire Sales Lift. Seasonality, competitor activity, distribution, advertising and product availability may change at the same time. A well-designed Controlled Experiment provides stronger evidence about Incrementality because it gives the business a more credible comparison for what might have happened without the intervention. The confidence of the conclusion should therefore match the strength of the measurement method.

The takeaway: Do not ask only how much you sold during the promotion, ask how much more you sold because of it

Promotional-period Sales are easy to see. They are not always the best measure of Promotion Effectiveness.
Before declaring a promotion successful, ask:

What was the Baseline?
How much Incremental Sales did we create?
Did the discount still generate Incremental Profit?
Did the growth come from New or Existing Customers?
Did we Cannibalize another product?
What happened after the promotion ended?

Answering these questions helps Marketing and Sales distinguish between a promotion that:
‍creates genuinely additional demand

and one that mainly:
‍discounts purchases that were likely to happen anyway.

That distinction is more useful for the next decision than knowing only that promotional Sales were “up 30%.”