Retailers have spent heavily on customer relationship management: loyalty platforms, customer data platforms, campaign tools and the analytics that connect them. The case is intuitive. Knowing who your customers are should help you sell more to them. Whether retail CRM actually does is harder to prove than most business cases suggest.
What a retail CRM programme actually costs
The software licence is often a small part of the total. A realistic cost base includes:
- Integration with POS, e-commerce and the loyalty engine, so that transactions link to a customer record.
- Data capture at the till, which adds time to checkout and depends on staff asking for a phone number at every transaction.
- Consent management and data protection obligations, which have become more demanding under India's data protection law.
- Messaging costs for SMS, WhatsApp and email campaigns.
- Staff who analyse the data and run campaigns.
- The cost of rewards and discounts given to customers who would have bought anyway.
The last item is the one most often ignored. A discount sent to a loyal customer who was going to visit next week anyway does not create a sale. It reduces the margin on one.
The attribution problem
Most CRM reporting measures response: how many customers who received a campaign later made a purchase. That number is almost always flattering, because the customers most likely to respond are the ones most likely to buy regardless. The useful measure is incremental revenue, the extra spending that would not have happened without the campaign.
The standard way to measure it is a holdout group. A randomly selected share of eligible customers receives no campaign, and their behaviour is compared with those who did. The difference, after deducting the cost of discounts and messages, is the real return. Retailers that do not run holdouts consistently cannot separate the effect of the programme from the behaviour of customers who were already loyal.
What a credible evaluation looks like
A retailer evaluating its CRM should be able to answer a short list of questions:
- What share of transactions is linked to an identified customer.
- Whether repeat purchase or visit frequency among identified customers has changed compared with a control.
- What incremental margin, not revenue, each major campaign type produces.
- What it would cost to achieve the same outcome through simpler means, such as better availability or sharper everyday prices.
A CRM that cannot answer these questions is not necessarily failing. It may simply not be measured. But spending should follow evidence, and unmeasured programmes tend to grow on habit.
The decision that matters is where the next budget goes: more campaigns, or the measurement discipline that shows which campaigns work. Retailers that fund holdout testing first usually find they can cut some campaigns entirely and spend more confidently on the rest.
Questions this story answers
How do you measure the ROI of a retail CRM programme?
Use a holdout group: a randomly selected share of eligible customers receives no campaign, and their behaviour is compared with those who did. The difference, after deducting the cost of discounts and messages, is the real return. Response rates alone are flattering.
What costs are included in a retail CRM programme?
Integration with POS, e-commerce and loyalty systems, data capture at the till, consent management and data protection obligations, SMS, WhatsApp and email messaging, analytics staff, and rewards given to customers who would have bought anyway. The licence is often a small part.
Why are CRM campaign response rates misleading?
Response counts customers who received a campaign and later bought. The customers most likely to respond are those most likely to buy regardless, so the figure is almost always flattering. Incremental revenue, spending that would not have happened without the campaign, is the useful measure.
