MARK1001 Chap.3 Repeat Buying and Segmentation
Repeat Buying and Segmentation
Repeat buying is analysed through purchase timing, purchase weight, penetration and average purchase frequency across a defined period. The Negative Binomial Distribution describes how buying is spread across many light buyers and relatively few heavy buyers.
This chapter separates buyer reach from buying rate, then asks whether segmentation variables genuinely predict a useful behavioural difference or merely create an attractive description of current brand users. Denominators carry the diagnosis: penetration concerns people in the market, whereas average purchase frequency concerns occasions among buyers.
Many infrequent buyers can matter collectively even though each contributes little, and a short window can make ordinary timing variation look like acquisition or churn. Segmentation is useful only when a measurable and reachable split predicts a difference large enough to change a real decision.
What this chapter covers
- 01
Purchase timing
- 02
Purchase weight
- 03
Penetration
- 04
Average purchase frequency
- 05
Negative Binomial Distribution
- 06
Segmentation bases
- 07
Brand-user profiles
Worked example · free
Reach explains the larger brand gap
- +2Compute Coast penetration as 18% and APF as 2.50.
- +2Compute Vale penetration as 32% and APF as 2.30.
- +1Identify buyer count as the larger difference.
- +1Recommend investigating reach and availability before frequency intensity.
Key terms
- Average purchase frequency
- Purchase occasions divided by buyers during a defined observation period.
- Penetration
- The proportion of the relevant population buying at least once during the period.
- Purchase weight
- How frequently different consumers buy from the category during a stated period.
- Segmentation basis
- A variable used to divide a market into groups for prediction and action.
Repeat Buying and Segmentation FAQ
Why do light buyers matter?
Light buyers contribute fewer occasions individually but are numerous. Their combined contribution can be substantial, and many apparent new or lapsed buyers are simply infrequent buyers moving in and out of short observation windows. A growth plan focused only on heavy users can miss the wider category base.
What does the NBD describe?
The Negative Binomial Distribution models how purchase frequency is distributed across buyers in a stationary or near-stationary market. It provides a benchmark with many non-buyers and light buyers and relatively few heavy buyers. It does not classify individual motives or emotional attachment.
When is a segment useful?
A segment is useful when the chosen variable predicts a meaningful difference, the group can be measured and reached, and separate action improves the decision enough to justify the cost. A vivid profile that does not change targeting, product or communication is descriptive rather than strategic.
Exam move
Recalculate penetration and APF for several invented brands, then explain which dimension creates the larger performance gap. Sketch a light-to-heavy buyer distribution and state why the observation window matters. For segmentation, take one demographic and one behavioural variable and write the evidence needed before either becomes a targeting rule. Start every numerical drill by writing the denominator.
Penetration divides buyers by the defined market population, while average purchase frequency divides purchase occasions by buyers. Use one invented table to calculate both measures for four brands, rank the penetration gaps and APF gaps, and explain which dimension contributes more to the sales difference. Keep percentage points distinct from purchase occasions.
Redraw the buyer distribution from memory with many non-buyers and light buyers and relatively few heavy buyers. Explain why the pattern is a market-level benchmark rather than a set of personality labels. Change the observation window from one month to one year and predict how apparent non-buying, repeat buying and lapse classifications may change.
Then state why a single table cannot prove stationarity or estimate an NBD model. Build a growth-diagnosis card with reach, rate, availability and time period. For any recommendation aimed at heavy users, ask how many buyers it can affect and whether the required frequency change is plausible. For any broad-reach recommendation, specify the physical or mental availability barrier it addresses.
This comparison prevents automatic loyalty-program answers. Practise segmentation as a decision test, not a persona exercise. For each candidate basis, record the behaviour to predict, incremental evidence, measurable group size, reachable channel and action that would differ. Reject a vivid segment if it cannot improve a product, targeting or communication choice.
Also write one overreach for each basis, such as treating age as motive or a past purchase pattern as a permanent identity. Finish with a timed case that combines metrics and segmentation. Award marks for correct denominators, computation, diagnosis, limitation and evidence-linked action. After answering, change one fact, such as distribution access or time period, and revise the conclusion.
If the recommendation does not respond to the changed evidence, it is probably generic rather than analytical.
Working through Repeat Buying and Segmentation in MARK1001? Sia is AskSia’s AI Marketing tutor — ask any MARK1001 Repeat Buying and Segmentation question and get a clear, step-by-step explanation grounded in how MARK1001 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.