Where Do You Draw the Line with Marketing Personalization?

98% of marketers agree personalisation helps advance customer relationships.*

The ability to tailor your marketing and CX communications based on what you know about a customer is generally a very good thing, but it can come back to bite you. We have put together a list of five things you should avoid when establishing your parameters for personalisation.


1.      When personalisation feels creepy

It’s really easy for personalisation to start to feel creepy, aka stalkery or big brothery, now that technology has given us the ability to match behaviour to individuals and use that to trigger messaging tailored to an individual.  A good example is when your outbound messaging happens too soon after a highly specific interaction. For instance, a customer has just had a conversation with a call centre or visited a branch or a store and performed a very specific (to them) transaction, then moves on with their day. Less than a minute after they have left the store or ended the call, they receive a text or email that encourages them to complete a survey about how their specific interaction went. While there’s nothing wrong with the concept of obtaining feedback about a customer’s experience while it is fresh in their mind (in fact there’s a lot right with it), acting too fast can feel creepy for the customer, like ‘big brother is watching’.

So if you’re planning to do something like this, wait at least a few hours after the interaction before sending the survey prompt.  Also, the more specific your communication is to the interaction, the more creepy it can seem.  For example, a message sent to a customer a few hours after they have purchased something that asks them about their general experience is a lot better than asking them why they didn’t purchase socks when they purchased 3 pairs of shoes 2 hours and 17 minutes ago.

2.      When personalisation talks to the wrong person

The next big mistake that’s made with personalised marketing is assuming that ‘one device or browser equals one individual’.  This fails to consider tablets and computers that are shared within a family, business, hostel or internet café.  It also tends to rely upon tracking methods, such as cookies, that are increasingly coming under more regulatory scrutiny, e.g. GDPR, and so are becoming less effective.  As it becomes easier to opt out of a cookie, the risk of inadvertently misusing them increases.

Consider the following example: an existing customer is looking at content on your website relating to a product they don’t already have, then they head out for a while.  Next time they visit your site, using the same device and browser, they might see a banner promoting a special offer related to the product that they were investigating and their status as a preferred customer.  If they’re using a shared device, the offer could end up in front of the next person who uses the device, and this person won’t qualify for the offer. Maybe a parent has been shopping for vehicle finance, then one of their under-18 kids (the next user) gets presented with a pre-approval offer because the cookie hasn’t been cleared.

For situations where you can’t be 100% certain your special offer will be presented to the right person, it’s a smart idea to only make offers to people who are logged into your site. Even then, with a shared device this logged in state can inadvertently apply to the next user of the device.  The safest and most effective way to get a specific offer to a customer based on online behaviour is to use a direct communications channel, such as email, where your chances of getting the right person are much higher.  Using a less direct channel, like website personalisation, for customer specific offers is just asking for trouble.

To address the issue of tracking becoming less effective over time, the best solution currently is to use indirect matching of behavioural segments to identified customer segments.  This involves sorting online visitors into segments by behaviour, then splitting up your identified customer base into the same segments using the few customers you are able to track online as models.  This will not be anywhere as effective as direct identification of individuals, but it will improve relevancy if the set of online visitors that you can identify is representative enough of your customer base.

3.      When personalisation is discriminatory

Facebook got in trouble when it let USA businesses target people based on certain demographic attributes. The most publicised example of this was allowing housing offers to be made to users of some races, but not others^. This contravened federal anti-discrimination parts of the 1968 Fair Housing Act.  Facebook has now agreed to make significant changes to its platform, so that advertisers can no longer target or exclude people based on characteristics like gender or race.^^

Rule of thumb: Always check your local laws before you present offers to a micro-targeted audience based on demographics.  You are generally on much safer ground if you use customer behaviour as a basis for your targeting, rather than demographics, but (as mentioned in point 2 above) you have to be aware of your local tracking and privacy laws.

4.      When personalisation is too adamant
Getting too forceful with an offer is another thing to avoid with personalisation. While tracking tools can tell you a lot about what a customer appears to like, acting like you know your customer better than they know themselves can be off-putting. Your tone shouldn’t become overconfident. ‘Gently gently catchy monkey’ is a smart sales strategy. Suggest and recommend, but don’t be too bossy.

5.      When individualised pricing makes you look bad

It’s important to remember the word-of-mouth factor if you have a strategy of using individualised pricing to reward loyal customers, rather than global discounts. If you want to present a price that reflects how you value a customer based on their buying habits, other customers can find out that you aren’t treating everyone equally and express their distaste on social media.  In the long term, this may not good for your brand and may also be in breach of local laws (see point 3 above).

When in doubt, call in the experts

At n3 Hub, we know all the rights and wrongs of personalisation. One thing we stress to our customers is that you mustn’t ever let the machines do it all. If you buy some personalisation software and use it without practitioner experience, it’s easy to make mistakes that can damage your reputation and upset customers.

Over the years, we’ve acquired a lot of expertise about implementing business rules to govern personalisation. We know that robust quality assurance must be applied to any automatic communication process.  Human judgement is an essential ingredient for a successful personalisation programme.

Talk to us about personalised automated marketing.

* https://www.evergage.com/wp-content/uploads/2018/04/Evergage-2018-Trends-in-Personalization-Survey.pdf

^ https://www.theguardian.com/technology/2019/mar/28/facebook-ads-housing-discrimination-charges-us-government-hud

^^ https://www.theverge.com/2018/8/21/17764480/facebook-ad-targeting-options-removal-housing-racial-discrimination