A common hot-button topic among companies looking to push to the next level is personalization. Whether it is ads, on-site, in-app, email, or anything else delivering a message customized to an individual is a powerful way to not only get their attention but also start a conversation. At the heart of any personalization, effort must be strong data that identifies unique characteristics about the people you are looking to personalize for. It used to be that generating this data was the most pressing matter for any company looking to do personalization but with advances in technologies like CDPs and first-party data collectors, the difficulty in offering personalization has started to shift away from data and towards content generation. The sheer volume of content production required to personalize a campaign or initiative can be overwhelming. Let’s look at a quick example:
If you are running a holiday sale and want to personalize it for your 5 main customer segments you essentially need to do everything 5 times. This could mean:
Content for website customization
– Landing pages
– Hero images
– Content blocks
– Popups and other limited offer modules
Ads
– Text ads
– Video ads
– Image ads
– Mixed media
Email campaigns
– Copy
– Imagery
– Links
Doing these things 1 or 2 times can be a large effort for many organizations let alone 5 times. We frequently see content development for personalized campaigns become such a bottleneck that limiting the scope is often required to keep things in the realm of possible. Fortunately, as with data issues prior, technology is once again poised to help alleviate some of the burdens of running a personalization initiative, this time with content. AI, in the form of technologies like Mutiny (https://www.mutinyhq.com/product), has started being tuned to produce content that is good enough to be usable in outward-facing touchpoints like website, email, ad copy or even certain types of imagery. That doesn’t mean a human shouldn’t review what is produced or fine-tune the language to match your brand’s tone, but it goes a long way in terms of getting something down that is 80% finished in very short order and it is this speed that is really key.
Personalization requires a lot of trial and error in the vast majority of cases. You may know your customer segments and even what types of offers or features those segments are excited about but nailing the exact verbiage can be tricky. A copywriter may be invaluable for writing the first version of something that the AI can work off of but sitting there generating dozens of versions of what they just wrote is typically a waste of their talents, which can usually be used elsewhere producing novel content. The variations being generated by the AI assistant would just need to be QA’d to ensure they meet brand standards and then used in whatever personalization efforts are in flight. If these efforts include A/B testing even better because the weaker variations will be weeded out quickly.
For now, the best way to prepare for these emerging technologies will be to establish a process to utilize what they produce – tons of stamped-out content. The typical approval process for content can often be burdensome, involving several teams and stakeholders. It makes sense, given that most content is destined to be external and everyone wants to maintain brand integrity but it also reduces the potential impact of having near-limitless content variations. Streamlining the approval process will be crucial in leveraging AI-assisted content production because of the sheer volume it can produce. Trying several versions of something and figuring out which is best will become more important than picking the perfect message out of the gate and the approval procedure should reflect this.
Going forward I would expect personalization engines to start integrating aspects of AI assistance directly into their product. Seeing as content production is often the limiting factor for getting a personalization effort off the ground alleviating this burden from users seems like a huge win for both the personalization industry and their clients.