Unless you’ve been living under a rock no doubt you’ve heard how great personalisation is for email marketing. Reignite exists because of the amazing outcomes our founders have experienced with personalised email. But increasingly the term personalisation is being used by vendors that covers a whole host of things, just so they can make it attractive to buyers looking to improve personalisation. At times it feels like personalisation is reduced to a tick box exercise. But the problem with that is not all personalisation is equal – dropping in a first name into an email I think most of us wouldn’t consider to add any incremental value to a campaign, but equally showing 2 different banners based upon gender isn’t going to deliver much compared to completely tailoring the product content to the customers tastes and past behaviours.
Unless you’ve been living under a rock no doubt you’ve heard how great personalisation is for email marketing. Reignite exists because of the amazing outcomes our founders have experienced with personalised email.
But increasingly the term personalisation is being used by vendors that covers a whole host of things, just so they can make it attractive to buyers looking to improve personalisation. At times it feels like personalisation is reduced to a tick box exercise.
But the problem with that is not all personalisation is equal – dropping in a first name into an email I think most of us wouldn’t consider to add any incremental value to a campaign, but equally showing 2 different banners based upon gender isn’t going to deliver much compared to completely tailoring the product content to the customers tastes and past behaviours.
Let’s look at what is and isn’t personalisation, and the various ways of implementing personalisation and their advantages and disadvantages.
My personal take on what true personalisation really is, is that it’s about content that is tailored to each individual’s tastes and preferences. It is here that the performance of personalisation can be observed at scale. The following is an example of an email that takes that individualised approach:
What isn’t personalisation?
First name and basic merging
I can’t be bothered to argue with anyone who thinks ‘Dear <Name>’ or merging in basic fields about someone is personalisation - it really isn’t meaningful personalisation so we shouldn’t include it.
This is sending different emails to different groups, and by it’s nature it is limited in how the content is tailored the content is to each individual. Let’s say you split your content by gender – while clearly sending ladies fashion items to women only and vice versa will work better than sending ladies fashion to men, it is still assuming all of those customers in the 2 segments have the same tastes in brands, styles, colours, price and more.
Often you’ll here of ‘personalised customer journeys’ by marketing automation vendors. What they mean 99% of the time is they fire triggered emails out at key events in the customer lifecycle such as new customers signing up. This really isn’t personalisation, it’s automation and segmentation combined. The main reason why these emails usually perform better than your business as usual mailings is these customers are naturally more likely to respond than the average customer, most of which are lapsed.
This often gets dressed up as personalisation when it is fetching content into the email automatically like the latest products from the website. In theory there can be personalisation involved here but very rarely is that the case, its normally generic content that if 2 people were to receive the email at the same time would get the same content. That ain’t personalisation.
What are the classic personalisation techniques?
This can also be known as conditional content but is generally if and else rules such as ‘If favourite brand = Nike, Show Nike banner’.
There are 2 usual problems with this approach. Firstly, it isn’t a particulary scalable approach as it takes time to write these rules limiting how many variants are practical. Even if you have the time to write say 20 rules, if you have to manually update the content behind them each time then this quickly becomes a burden the gets dropped.
The other typical problem is most ESP’s require conditional content to be implemented with a scripting language. That means that technical skills are required to implement rather than having a simple UI that anyone in marketing can use to create the variants.
If you have a mid-market or enterprise ESP (Not say a Mailchimp or ESP where the pricing is publicly available on their website) you might have options for storing content or other types of data in the platform that can then be used to merge into emails.
There are many different approaches to this. Some mature platforms allow the creation of a relational database in their platform, others will allow you to connect a data feed to your template where you can suck in content at the point of send.
The advantage is you can sale to 1:1 personalised content with each individual getting content unique to them.
But both these options rely on you doing most of the logic outside of the platform requiring lots of data resource, and then coders to transform it into content. This slows down the ability to both implement but also evolve
Probably the most hyped and mis-understood technology coming to email marketing is Artificial Intelligence (AI). AI is often positioned as being naturally superior to other approaches because a computer is making the decisions.
In email marketing it is used to predict the best time to send an email, what customers are likely to unsubscribe and what we want to talk about here, recommending products.
The key word we used above those is predict – it isn’t putting the best products in the email, it is making a prediction. Predictions are only as good as the amount of data you have to support them, and the quality of the logic you have used to crunch that data.
And this is where so many AI product recommendations for email fall down. For example many of these will use the same set of data points and rules if you were selling flights, a discount homewares store or a luxury fashion retailer. It’s like a horse racing pundit using their prediction methods to guess share prices.
Without control of these not only do you get less accurate results, you also can’t add your own merchandising requirements to tailor these for your specific business objectives like clearing seasonal stock or boosting new product ranges.
With AI for email its very much a case of not falling for the hype and spending time critically appraising any solutions.
Taking all of the above into account you’d be left thinking there isn’t a lot going for email personalisation. That would be a bit harsh as for many of the smaller use cases they are perfectly adequate – but they fall down when you want to transform your email programme with personalised content at the heart of it.
Our approach is based upon 20 years experience of doing it the old way and suffering the limitations we’ve built a recipe builder that allows you to build your own personalisation algorithms without the need for marketers to get involved, and having full control to customise as required to fit your specific needs. We won’t go on about it here because you can see more on our site or in this video, but you don’t have to be limited to traditional methods that have become almost standard in the email industry.
It’s important to think about how you will implement personalisation is because it will have a direct impact on whether you achieve your goals. If you only have methods that take too much manual resource you won’t ever scale personalisation. If you choose technology that isn’t accurate why on earth will that deliver any return. And if you have something you can’t control you’ll quickly find it difficult to integrate personalisation into your existing campaigns.