From siloed data to one of South Africa’s most influential loyalty programmes

A big four bank wanted to reward customers for banking better. It had the ambition and the customer base, but its data sat in departmental silos and nobody could say what behaviour change would actually be worth.

The challenge 

One of South Africa’s big four banks wanted to bring a new loyalty programme to market. The intent was specific: reward customers for positive changes in their banking behaviour, and for spending with the bank’s retail partners. 

The behaviour the bank wanted to encourage was clear. Almost nothing else was. 

Individual customer data was spread through different departments of the bank, with no consolidated Single View of Customer to work from. That made even basic questions hard to answer. Which customers were worth incentivising? How much would it take to move any one of them? And if the behaviour did change, what would that be worth to the bank? 

Those three uncertainties, how to incentivise, how to target, and what it does to the bottom line, are where loyalty business cases usually fail. A programme gets designed around what the business would like customers to do, launches, and then cannot prove its own value. The bank was determined not to build that. 

What we did 

We started with the data rather than the design. 

1. A Single View of Customer. We extracted historical data at customer and transaction level, across multiple products, and consolidated it into a single view. For the first time the bank could see a customer whole, rather than as a set of departmental fragments.
2. Longitudinal customer analysis. With that view in place we analysed how customers actually behaved over time, and used it to shape the programme rules and the customer value proposition. The proposition was built to fit observed behaviour rather than assumed behaviour, which is a materially different starting point.
3. A dynamic financial forecasting model. We constructed a forecasting model to support the business case, so the bank could test multiple design scenarios and see what each would cost and return before committing to any of them.
4. A data warehouse and analytics environment. We designed and built the environment that would run the programme after launch, covering monthly reporting, campaigning, and uplift calculation of the value added to the bank. The measurement was designed in from the start rather than retrofitted. 

The result 

The bank went to market with a customer value proposition suited to how its customers actually behaved, and a detailed assessment of the programme’s costs and profitability across multiple design scenarios. 

Once launched, the programme was rated: 

• Retail bank programme with the most influence on where customers shop 

• One of the top three programmes consumers cannot live without 

Eighty20 did not hand the programme over and leave. We remain the bank’s analytics partner, providing ongoing analytics, campaigning and strategic consulting to the programme we helped design. 

“Their team’s expertise, professionalism and collaborative approach has made them a trusted partner.”

Loyalty and Payments Lead, Astron Energy

A client reference from Eighty20’s wider loyalty portfolio, not from the bank described in this case study. 

Why it worked 

Loyalty programmes are usually designed by marketing and measured, if at all, by finance, some time later. This one was designed and measured by the same team, from the same data, in the same environment. 

That is why the business case survived contact with reality. The forecasting model, the warehouse and the uplift calculation were built before launch, so the bank could answer the question every loyalty programme eventually gets asked: what did this actually earn us?

Thinking about a loyalty programme, or trying to prove the value of one you already run? Talk to us about building the customer view and the business case together.