Finding R5 billion of growth in a customer base the client already had

A large retailer with several million active customers was growing more slowly and competing harder. It had a single view of its customers and was not using it. Four analytical steps later, it had a marketing roadmap against a quantified growth opportunity.

The challenge 

Our client, a large retailer with several million active customers, was experiencing slowing growth and increasing competition. In that environment it needed to settle two things: what its growth strategy should be, and how it should be positioned in the market to deliver on it.

The retailer was not short of data. It already had a single view of its known customers. What it did not have was any way of leveraging those insights across the organisation, so the insight sat in one team while range, location, pricing and marketing decisions were made in others.

Eighty20 recommended a piece of work to determine the client’s key target consumer segments, identify and size the market opportunities available to it, and then develop a set of strategies and marketing priorities to capture that growth, covering customer engagement, proposition design and brand positioning.

What we did 

Working with the client and its advertising agency, we ran the GrowthMap framework in four steps. Each step answers one question, and each one narrows the field for the next.

The four steps of the GrowthMap framework, from external market view through to an execution plan.

1. External view: who is out there, and where are we winning?

With appropriate consents, we used unique identifiers already held by the client, such as SA ID number and cell number, to match its entire customer base to the Eighty20 National Segmentation: 1,500 microsegments, 46 sub-segments and 8 main segments.

That match enriched the client’s customer view with thousands of variables, ranging from demographics through to retail, financial, digital and media consumption behaviour.

Match completion. 84% of records matched to the ENS, and 79% of matched customers matched at high confidence.

With the base matched, we could determine the client’s penetration into the market segment by segment. Critically, the client could now see the profile of its customers and of its not-yet-customers, and where it was performing well or poorly against each.

Penetration by ENS segment. The client was well penetrated into more affluent segments and had limited penetration into Comfortable Retirees and Students and Scholars. Mass Market showed low penetration, and therefore opportunity.

2. Internal view: how are our own customers behaving?

Overlaying internal transactional data onto the enriched view let us understand how different customers were actually engaging with the client: who the top customers were and what made them different, and who was moving closer or further away, and why.

Value driver trees. We used value driver trees to decompose the drivers of sales, average transaction value, price per unit, number of customers and so on, by segment, by category and over time. One finding set the tone for the whole engagement: although sales were up strongly, the growth was coming from acquisition, while existing customers were spending more often but with smaller baskets.

Value driver tree by ENS segment. Decomposing total sales into customers, frequency, transaction value and price per unit exposed that customers were buying cheaper products across all segments.

Attribution analysis. We then used data science techniques to quantify exactly which drivers were contributing to growth and which to value erosion. The answer was uncomfortable and useful: acquisition in the Mass, Middle and Heavy Hitter segments was driving growth, while the buying down of those same segments was the single largest contributor to value erosion.

Attribution analysis: the top and bottom ten drivers by impact on total spend, separating what was creating value from what was eroding it.

3. Opportunity view: how big is the prize, and where is it?

The first two steps describe the world. The third one sizes it. We ran a four-part analysis:

1. Score every customer. We performed look-a-like analysis of customers within each microsegment, scoring each customer on their loyalty or current spend potential relative to people who look like them.

Look-a-like analysis. Customers are grouped by age, income, marital status and gender, and the distribution of spend within each group establishes the maximum potential for the current business model.

1. Distribute the base. Distributing all customers according to that spend potential reveals the actual level of loyalty across the customer base, rather than the level the business assumes it has.

Distribution of customers by share of wallet, from minimal through to high share of wallet customers.

1. Size the nudge. Assuming existing customers can be nudged into becoming slightly more loyal, we identified and sized the growth opportunity available from each customer sub-segment. This is where the opportunity stops being a theory and becomes a number.

Total opportunity value by sub-segment and share-of-wallet group, with the point at which 80% of the total opportunity is accounted for.

1. Make it actionable. Because the analysis runs at customer level, each individual customer can be identified and profiled with their own targeted growth potential. That is what makes the output usable by a campaign team rather than only by a strategy team.

Opportunity plotted against average income and average age by ENS sub-segment, with bubble size showing the number of customers.

4. Marketing priorities: what do we actually do on Monday?

With the customers understood and the opportunities sized, we recommended a growth strategy and marketing roadmap to acquire more than R5bn in growth.

Where growth will come from, split between new customers and existing customers, with the sub-segments carrying most of each opportunity.

The priorities that came out of it were specific:

• Grow the customer base by targeting Middle Class Workers, emphasising value products in specific categories
• Increase frequency among existing Heavy Hitters by emphasising everyday consumables
• Use the detailed profiling operationally, to drive location strategy, product ranging, promotional strategy and hyper-personalisation

Opportunity sized per priority segment, split between the client’s existing customers and consumers who are not yet customers.

Opportunity sized per priority segment, split between the client’s existing customers and consumers who are not yet customers.

Because every priority is anchored to identified, profiled customers, the roadmap could be handed to the teams that make range, location, promotion and communication decisions, which was the gap the client started with.

The result 

• A recommended growth strategy and marketing roadmap against more than R5bn in identified growth opportunities
• A quantified view of penetration and opportunity across 8 segments, 46 sub-segments and 1,500 microsegments
• 84% of the customer base matched to the ENS, 79% of matches at high confidence
• Customer-level growth potential that could be used directly for ranging, location, promotion and personalisation 

“Acquisition in the Mass, Middle and Heavy Hitter segments was driving growth. The buying down of those same segments was the largest single contributor to value erosion.”

Why it worked 

Growth strategies usually fail on one of two counts. Either they are directionally sensible but unsized, so nobody can prioritise between them, or they are sized but not actionable, so nobody can execute them.

GrowthMap avoids both by running the same analysis all the way down. The external view sizes the market, the internal view explains the current trajectory, the opportunity view attaches a rand value to each sub-segment, and because it all resolves to individual customers, the strategy arrives with a target list attached.

Slowing growth, and a customer base you suspect you are under-using? Talk to us about running a GrowthMap on your own data.