The TRB Share Prediction Model

Numerous studies on new product launches across a range of FMCG categories including personal wash, fabric cleaning, toothpaste, hot beverage categories and infant milk, were conducted by the author with Hindustan Unilever’s consumer panel data in 1990 - 94, and more recently with FairPrice supermarket’s loyalty card data. Market share prediction for the new products was based on the notion of trial and repeat purchase that forms the basis of the Parfitt-Collins model as well as BASES and other simulated test market models. The outcome was a refined version of the Parfitt-Collins technique, which I simply called the TRB model.

The TRB model is a consumer/loyalty panel based technique for predicting a new product’s expected market share soon after it is launched.

Market share is a variable, and in an ever changing environment it constantly fluctuates. These fluctuations are more pronounced particularly during the launch of a product when consumers are trying out the product, and manufacturers are strongly promoting it. It usually takes more than a year for the sales baseline to stabilize, fluctuating over a narrower bandwidth. The TRB model’s share prediction pertains to new product’s expected market share when this relatively stable state occurs.

By decomposing a new product’s sales in terms of fundamental growth drivers — trial and repeat purchase — this model is able to estimate the expected share of the new product. These drivers which reveal consumers’ desire to try the brand and their willingness to continue buying after experiencing it reflect the brand’s ability to succeed at the FMOT and the SMOT.

Trial and Repeat Purchase

To be adopted by consumers, a product must succeed at both the first and the second moments of truth. It can succeed only if it develops a substantial base of regular consumers who are drawn to try the brand, and having tried it, are induced to continue to buy it on an ongoing basis.

Success at the FMOT can be gauged by the trial rate, the proportion of consumers who try the new product. It reflects the attractiveness of the new product concept, how effectively it is communicated, and how well the brand is positioning. The trial rate is also strongly influenced by causal factors such as distribution, promotions and sampling.

Success at the SMOT depends on whether consumers’ experiences with the product evoke the desire to continue buying it. It is a function of the extent to which the new product lives up to or exceeds expectations.

Trial Index

$$ Trial\,Index = \frac{Trial\,Rate\,(Product)}{Trial\,Rate\,(Category)} $$

Where:
Trial Rate (Product) = Projected Penetration (Cumulative) of new product.
Trial Rate (Category) = Projected Penetration (Cumulative) of category over the same period.

The trial index reflects the first moment of truth and is dependent on:

  • The product concept (inclusive of price) and its communication through advertising and packaging
  • Product distribution.
  • Consumer promotions, including sampling.

Repeat Buying Rate (RBR)

The RBR is a measure of the propensity of consumers to continue buying a product. In terms of definition, RBR(t) is the brand’s share among those who repeat purchased the brand t periods (usually the time period is in months) after they first tried it. RBR (5) = 15% means that for the average trialist, the new product constitutes 15% of purchases of the product category, on the 5th interval after trial.


Exhibit 11.15   Illustration of RBR: Purchases by three consumers on the fifth month after they first tried the brand.

Exhibit 11.15 illustrates how RBR is computed. It depicts three trialists of a new product: Anita, Betty and Claire. Anita bought the new product in January, Betty in March and Claire in April. The fifth interval after trial is June for Anita, August for Betty and September for Claire. On their respective fifth interval, Anita bought 10 units of the new product and 20 units of the category as a whole, Betty bought zero units of the new product and 60 units of the category, and Claire bought 5 units of the new product and 20 units of the category. Their total purchases of the new product (10 + 0 + 5) as a proportion of the total purchases of category (20 + 60 + 20) equals 15%, which by definition is RBR(5).

Note that RBR computation is based on all those who ever tried the new product, which includes current as well as lapsed buyers. It is measured over each purchase occasion after trial — so there exists a whole series of RBRs. RBR(1) is the RBR value at first purchase interval after trial, RBR(2) pertains to the 2nd purchase interval, RBR(3) to the 3rd and so on.

As novelty wears off, the level of interest in a new product is likely to diminish. Promotions and advertising which peak during launch, decline thereafter. With fatigue setting in, RBRs tend to decline before they stabilize. RBR(1) is usually higher than the other RBRs, followed by RBR(2), RBR(3) and so on. Yet the rate of decline rapidly decelerates and RBR values stabilize after the first few intervals, for successful new products. This stable value of the RBR is a measure of consumers’ willingness to continue buying a brand. It is the parameter in the TRB model that usually has the greatest bearing on the success of the new product.

Buying Index

Trial × RBR is a fair approximation of a product’s expected market share, except that it assumes that the consumers buying the new product tend to buy as much of the category as the average category buyer. To factor the heaviness of buying of the category by consumers of the new product, the TRB model introduces a third parameter, the buying index:

$$ Buying\,Index = \frac{Category\,Consumption\,by\,Brand\,Repeaters}{Category\,Consumption\,by\,All\,Category\,Buyers} $$

Where Category Consumption is the average category volume purchased per buyer.

The buying index is a measure of the heaviness of buying of the product category by the new brand’s repeat buyers, and is computed as the ratio of purchases of category by the brand’s repeat buyers over the purchases of category by all category buyers. If, for instance, the buying index is 1.25, it tells us that the average repeat buyer of the brand buys 25% more of the category than the average category buyer.

The Model

Market Share = Trial Index × RBR × Buying Index

Provided the sample size is adequate, the above forecast provides a reliable estimate of the brand’s baseline market share, till the time market dynamics change. Since dynamics do tend to change, if panel data is readily accessible, marketers should continue to periodically revise estimates of a new product’s market share over the first year or two of launch.

The TRB model works best for those categories where the inter-purchase interval is small. Categories like bread and fresh milk for instance, that are purchased at least once a week, can be forecasted within a few weeks of launch. The model also works well for products bought on a monthly or quarterly basis, which includes most packaged foods and personal care products. It is not, however, recommended for products that are infrequently re-purchased.

Diagnostics


Exhibit 11.16   Trial, RBR and sales patterns.

The underlying patterns for the build-up of trial, RBR and sales are depicted in Exhibit 11.16. The sales trends vary depending on the RBR and the pace of growth of penetration, and there is often a kink or a hump during the first year of launch. The existence of these diverse patterns that may trend up and down explains why early sales data does not provide a good reflection of the share the new product is likely to achieve. Moreover, sales data is also affected by causal factors like promotions and advertising.

RBR has a lasting impact and is usually the prime determinant of the long term success of a product. If a product has low RBR, marketers need to investigate, via customized research, the reasons why consumers do not repeat purchase. Raising RBR usually requires refining some of the elements of the marketing mix.

If RBR meets or exceeds expectations but sales are constrained due to low trial rate, marketers may use sampling, price discounts, product displays and other forms of promotions to induce trial. They also need to examine the product proposition and the manner it is communicated, to understand why target consumers do not find the product appealing enough to try it.

Though well-executed promotions yield valuable gains in the trial rate at a time when the new product is striving to gain adoption, their impact in the long run is diluted due to lowering of RBR. The propensity of consumers induced to try the new product through promotions to continue buying the brand is likely to be lower than that for those consumers who were drawn to it, irrespective of the promotion.

Marketers should also use the gain–loss analysis to determine the source of growth of the new product. Of interest is the extent it is cannibalizing the company’s other products.

Previous     Next






TRB Share Prediction Model
TRB Share Prediction Model

TRB Model for predicting Market Share