Did your promotion create demand, or just move it forward?

The promotion ends with an impressive sales chart. Orders rose, the event beat its target, and the commercial team wants to repeat it. Then the following weeks soften. A replenishment order placed during the spike arrives just as demand slows, and finance asks whether the discount funded growth or subsidized purchases customers would have made anyway.

The event-week result is not necessarily wrong. It is incomplete. A promotion can attract additional buyers, accelerate planned purchases, move demand between products or channels, or change the quantity purchased per order. More than one effect can happen at once.

The useful business question is therefore broader than "did sales rise?" It is whether the event supported its intended objective over a suitable period, after considering timing, economics, availability, and the strength of the evidence. That review can begin with ordinary retail data, provided the team is careful about what it can and cannot conclude.

Give the promotion one primary job

A promotion intended to introduce a product should not automatically be judged like a clearance event. A customer-acquisition offer, a retailer-supported feature, and a stock-reduction campaign have different reasons to exist. Define the primary objective and the limits before the event starts.

For example, a team may seek to reduce aging inventory while maintaining a minimum contribution. Another may want new households to try a category, with a later review of repeat purchase. A third may support a retailer commitment where availability and execution matter alongside direct economics.

Record secondary outcomes, but do not change the definition of success after seeing the results. If an event designed for contribution growth misses that objective, an increase in traffic may be useful learning without making the original plan successful.

Look before, during, and after

Consider a simplified illustrative product that sells 1,000 units in each comparable period without an event. The promotion period records 1,600 units, suggesting a 600-unit increase. The following comparable period records 500 units. Across the event and following period together, the observed total is 2,100 rather than the 2,000 implied by that simple baseline.

That arithmetic raises a useful question about purchase timing, but it does not prove that only 100 units were incremental or that the event caused the decline. Seasonality, stockouts, competing offers, distribution, and the quality of the baseline could change the interpretation. Treat the comparison as a diagnostic starting point.

Choose a review horizon that fits the category. A frequently replenished item and an infrequently purchased durable product should not inherit the same observation window. Also check for a pre-event dip: customers who expect an upcoming offer may postpone purchases, making the event-week increase look larger.

Distinguish new demand from several kinds of movement

Ask whether the event changed the total business result or only where and when the result appeared. Relevant movements include:

• Timing: customers buy earlier, later, or in larger quantities than usual.
• Product substitution: the promoted item gains while a related item loses.
• Channel switching: one retailer or storefront gains business that moved from another.
• Customer mix: existing customers account for the increase while the intended new audience changes little.

The evidence needed differs. Basket detail can help examine product relationships. Comparable channel coverage can reveal redistribution. Reliable customer history can distinguish recorded new and returning buyers under an agreed definition. Missing data should narrow the conclusion, not be interpreted as an absence of displacement.

Do not assume these effects are inherently bad. Moving demand can be valuable when the objective is cash recovery or stock reduction. The problem is calling every movement additional growth and then building the next budget or replenishment plan on that assumption.

Bring margin and event costs into the review

Revenue growth can hide a weaker economic result when discount depth, product mix, returns, or fulfillment costs change. Compare contribution under one documented cost definition, not just gross sales. Show which costs are included and which remain unavailable.

Event costs may include retail funding, promotional fees, creative production, incremental advertising, and other agreed expenses. Take care not to count the same discount or funding item twice. Some costs are embedded in the sales or margin data; others need a separate adjustment.

Where returns arrive later, report an early economic view as provisional. A fast report can support an operating decision while a later view closes the event. Keep both versions and explain why the conclusion changed. This is more credible than silently revising an earlier claim of success.

Do not confuse a weak comparison with a controlled test

A previous period or an unpromoted product can provide context, but neither automatically shows what would have happened without the promotion. The comparison might differ in availability, season, price, distribution, or customer demand.

Where feasible, design a suitable holdout or other credible evaluation before launch. Google's conversion-lift documentation illustrates the distinction between experiment-based incremental outcomes and standard attributed results. That specific advertising method is not a universal promotion test; the broader lesson is to state what the comparison can establish and respect the design's assumptions.

If a controlled design is impractical, use transparent baseline estimates and sensitivity checks. Compare more than one reasonable reference period, identify concurrent events, and show whether the conclusion changes. A conclusion that depends on one convenient baseline deserves less confidence than a pattern that survives several defensible checks.

Include the supply consequences

The promotion result affects more than the next marketing calendar. An event can drain stock from high-value locations, create fulfillment pressure, or make an ordinary demand forecast look like a permanent step change. Review receipts, inventory snapshots, service interruptions, and any product availability gaps beside sales.

After the event, separate temporary promotional demand from the ongoing replenishment assumption. Avoid mechanically carrying the peak into the next order. Equally, do not interpret a post-event decline as a permanent demand collapse without checking whether shoppers stocked up, stores ran out, or distribution changed.

Commercial, planning, and finance teams should agree which observations change the next action. That may mean adjusting discount depth, moving an event date, reducing purchase quantities, or selecting a different product group. A review is useful when it changes a decision, not merely when it produces a polished retrospective.

Where UJAS fits

UJAS Business Intelligence brings this promotion review into a governed retail view. Where promotion-event analysis is configured, teams can examine linked sales, margin, event costs, pre-period comparisons, related-product context, and post-event behavior together. The analysis depends on the event setup, linked product scope, and available source data. Baseline estimates help investigate performance; they are not experimental proof of additional demand caused by the event.

This gives commercial, planning, and finance teams a shared starting point: which products were promoted, how the result compares with the selected baseline, and what happened afterward. Follow-up questions can then focus on the proposed next event rather than rebuilding the same numbers in separate spreadsheets. For Amazon-focused decisions, Skynar Commerce Intelligence is designed to connect advertising with inventory and economic context; the available capabilities are confirmed for each engagement.

Start with the most recent event whose success is still debated. Write its original objective, review the periods around it, and reconcile the economic assumptions. Read our related guide to forecast evaluation and planning decisions, or talk with VeriVal about building a repeatable commercial review.

References

• Google Ads: incremental outcomes versus attributed conversions
• Google Ads: geographic conversion-lift study design and eligibility