Retailers today find themselves in a tricky balancing act. On one hand, customers are more demanding than ever. They not only expect great experiences, but they expect brands and retailers to tailor those experiences to their unique needs and desires. According to the Salesforce “State of the Connected Customer” report, 70% of customers say understanding how they used products and services is important to winning their business. And 59% of customers say tailored engagement based on past interactions is very important to them. It is safe to say retail personalization is here to stay.

At the same time, companies find the spotlight shining on their handling of customer data is brighter than ever. Privacy is a sensitive topic, and companies that are perceived as stepping out of bounds are punished quickly.

We have shared before (Don’t Stand So Close to Me) about the perception that personalization can be creepy when it is not done well. As the chart below shows, customers who experience an interaction as creepy will not be shy about taking action based on their negative experience.

Source: InMoment 2018 CX Trends Report

Get the basics right

If you are handling customer data, you need to handle it in a way that earns their trust. Retailers, as well as any partners involved in personalized marketing, need to follow the standards. In the European Union, this means GDPR compliance. In other parts of the world, a similar standard should be met…whether or not it is required by law. Customers should have access to clear information about how their data is being used, and to an opt-out process that is straightforward and easy to find. Wherever possible, the systems involved in generating personalized recommendations should not even have access to personally identifiable information, such as names, social security numbers or email addresses.

GDPR compliance is the starting point for customer privacy, but it is not the end. None of this will set you apart. It will only get you in the game. Keep reading to learn how the best retailers earn – and keep – the trust of their customers.

Make the right mistakes

The algorithms used to determine the right offer for the right customer are getting better all the time. But the nature of the game is that sometimes you will still get it wrong. There is no system that can give every customer the perfect offer every time. The key is to be smart about which mistakes you are willing to make…and which ones you need to avoid.

Customers will certainly forgive the occasional miss if a personalized offer is not on target. If a retailer offers Corn Flakes and a customer prefers Rice Krispies, there is no harm done. However, other offer misses can create a more negative customer experience. Think of a Muslim receiving an offer for alcohol, or an Orthodox Jewish customer receiving an offer for pork.

Your personalization tool should allow you to flag items that should only be offered if a customer has purchased them before.

Avoid the past purchase trap

Sending offers on something a customer has actually purchased before is safer, but even here danger lurks. There are plenty of scenarios where a customer may have purchased something in the past but it is still not relevant for them. Think of someone who purchased baby products as a gift for a baby shower, or someone who bought meatless burgers because they were hosting a vegetarian relative for the week.  Well-designed, self-learning personalization algorithms can help immensely here. A good personalization tool will also take into account the total number of past purchases as well as purchase recency to ensure the customer is still interested in the product.

Even if a customer is clearly a regular purchaser of a product, they may still not want to get offers on it. Items like pregnancy tests or weight loss products simply feel different to people than crackers or pasta. For certain categories, it is best to avoid offers altogether. Usually, the potential downside of sending offers like these far outweighs the additional sales they might generate.

Practice active listening

Even if an offer is not flagged as sensitive, sometimes it will just be the wrong item for a particular customer. Personalization systems should have built in controls that stop sending an offer after it is ignored a certain number of times. Even better, put the customer in the driver’s seat and let him tell you when he no longer wants to see offers for a product. If he is not going to but it anyway, there is no risk to removing it from his rotation. And by respecting his wishes, you make it much more likely he will engage with your personalization program going forward.

When it comes to respecting customer privacy, it is in everyone’s interest to go beyond the minimum required by law. Follow the principles above and you will find yourself leading the pack.