There has been a spate of articles recently about the value of using granular data to target customers more accurately and personalize product recommendations (example 1, example 2). The tone of the articles has been positive, bordering on breathless. And given CQuotient’s raison d’être, we are only too happy to see the coverage.
That said, the coverage is a bit imbalanced. Personalization, as we have come to understand from many retailer conversations, is not without risk.
Contrast an individualized email campaign featuring personalized product recommendations with a batch-and-blast campaign that shows the same “Our most popular items” list to every customer.
The individualized campaign may either elicit a “Wow, they really get me!” reaction or “”What the heck were they thinking?” reaction.
The batch-and-blast, on the other hand, is likely to affront no customer. Of course, no customer is likely to be delighted either.
If you get the personalization right, you will delight the customer, move the needle, and be a hero. If you get it wrong, you may get the dreaded visit from a senior executive asking why their spouse was recommended item X.
If you continue to batch-and-blast, you won’t rock the boat, you will maintain the status quo. You won’t get fired. Chances are you won’t get to become a hero either.
So what’s a marketer to do?
Being relevant is not an all-or-nothing proposition. Start small. Measure the results. Scale.
- Choose a small set of customers to individualize first
- Add a small personalization block to your batch-and-blast template
- If you email 3 times a week, start to personalize one of those 3 emails
Personalization doesn’t automatically mean individualized product recommendations. Better matching of versions to customers is a good example. If you have a few different versions or themes for your emails, match each customer to the theme that best speaks to them. If you have a great idea for a new campaign (30% off Justin Bieber tees!), leverage technology to target just the right audience for the campaign rather than doing a crude select.
Keep in mind that there are plenty of things that can be done algorithmically to minimize the risk. Ask your algorithms partner what safeguards they have in place.
No technology that promises to delight your customers and move the needle for your business is without risk.
The real question is: Which is riskier? Trying it or not trying it at all?
Amazon attributes 30% of its $48 billion in revenues to its recommendation technology. Do you think your competitors haven’t noticed?