The Real Challenges of Email Personalisation

We've all seen the vendor-backed surveys: glossy reports that confidently tell us why brands are struggling with CRM personalisation.

They present familiar charts and familiar answers - data siloes, data quality, lack of resources - each ready-made to support the vendor's pitch.

But in practice, those aren't the real blockers anymore.

If you're in CRM or marketing operations today, chances are:

• You already have your key data sources integrated.
  
  
• Your product and transactional data is clean enough to power your ecommerce engine.
  
  
• And no, you're not short on resources - you're short on impactful execution.

So what are the real reasons personalisation in email CRM is still so hard?

1. Data Breadth: Most Customers Are Strangers

Even with all your integrations in place, your data may still be shallow. If 80% of your customer base has only transacted once, that's a thin foundation to build personalisation on. What do you really know about them?

Zero-party data collection sounds great in theory, but in practice it's often sparse, fragmented, and hard to action.

This is where predictive solutions have real value. By analysing what limited purchase behaviour you do have - along with customer attributes - they can model likely affinities for categories, brands, or styles. In essence, they give you something to work with when your data lake is more like a puddle.

2. Data Transformation: From Raw to Useful

Even with good data, too often it's just sitting there - raw and unstructured in your ESP or CDP.

Having the entire raw transactional history doesn't tell you what the customers favourite brand is, what the last thing they bought (and therefore what next).

True personalisation requires investing time (and often tech) to transform raw data into actionable traits.

At a really simple level that might be a series of fields exposing what the last purchase involved - brand, date, category and so on - or what their most popular attributes are. Beyond that it's building models that predict affinity to other parts of your catalog based upon what other customers also did.

Without these your personalisation will be in one of two camps:

• Lots of manual crafted small, segmented campaigns or horrendously complicated if statements that won't scale
  
  OR
  
  
• Plonking in a generic AI product recommendations widget into an email and hoping it will be noticed and perform (neither will happen)

Don't think data is just a transfer and synchronization task - it's way more than that.

3. Orchestration: Content ≠ Personalisation

Let's say you've done the hard work: data's integrated, customer traits are calculated, and you're pulling in relevant product content. Now what?

You're still likely hitting a wall when it comes to orchestration - you can't just send the same content every single time, you need the personalisation to adapt on each send, deciding what goes where, in what order, and how to avoid fatigue.

If we take the example of personalising the product selection most out-of-the-box product personalisation tools for this are limited as you look deeper.

• They tend to show the same recommendations in each email causing fatigue and limit usage
  
  
• They can show a narrow part of your catalog - often as closely related to past items as possible. That's fine if you are trying to convert them in an online session, but rubbish if you are trying to inspire their next purchase months later
  
  
• Results look like a computer has picked them, not crafted by a visual merchandiser. Everything ends up being the same dull colours or too many of the items come from the same category

I can't help but think many vendors have either repurposed many built for web personalisation tools, or simply stopped after building the overall personalisation capability without thinking about how one email campaign doesn't exist in isolation, but it's part of a wider series of touchpoints. Without considering this the solutions can't be used in as many campaigns and therefore don't deliver the scale needed..

You'll find a few more modern solutions are coming out where you'll see features to address these specifically for email, namely capabilities they might call ‘guardrails'. Look beyond the basic personalisation capability.

4. Expertise: The Personalisation Experience Gap

Here's an uncomfortable truth: very few people have actually run successful, scaled CRM personalisation projects. If this is your first time attempting one, chances are high you'll hit pitfalls that aren't in the playbook.

And if the first test underperforms, personalisation quickly becomes the scapegoat.

The best teams either have someone who's done it before - or they're humble enough to bring in outside expertise. Without that, you risk months of effort for little to no uplift, and it becomes harder to justify another try.

5. Commercial Pressure: Reality vs the Personalisation Dream

We all want to get to that CRM nirvana where every email is tailored, every message timely. But real-world marketing teams are under constant pressure to hit targets, move stock, and drive immediate revenue.

Face facts - you aren't going to replace all of those promo codes with personalisation, they generate too much revenue that keep your boss's boss in a job for another quarter.

But personalisation can boost their performance, but also help deliver some other strategic goals. Here are a couple of examples.

• A holiday company was blasting their database with ‘Late Deals' - if they didn't sell the remaining spots they would lose money on the tours. The trouble was if you didn't live in a 10 mile radius of the pickup points you were not going to book, and they couldn't promote all the late deals as there wasn't enough space in the calendar. Personalisation solved this immediately, matching the deals to the right customers - and a 312% improvement was gained
  
  
• Have a bunch of high margin product categories that only appeal to certain customers? We worked with a TV shopping company who really wanted to push these items for obvious reasons. But blasting wasn't the solution, nor simply relying on those who had bought them before - it needed more scale. Thanks to predicting category affinity they were able to find the types of customers in the database most likely to buy
  
  
• Years ago in my footwear days we had this very thing with a subset of customers who were in a comfort segment. They'd buy all of their footwear from us because of health issues such as diabetes that some of our ranges seemed to be popular with. But the majority of our customers were ‘fashion' buyers so our main email programme ignored them. Introducing some simple semi-automated campaigns to what we called this ‘underserved segment' helped us grow revenue from these customers for little effort.

If personalisation doesn't serve the business needs of the moment, it won't be prioritised - find additional ways it can solve these business problems for getting it on the roadmap.

In Summary

Email personalisation isn't failing because of the tired tropes we see in vendor slides - data siloes, bad quality, or resource shortages. It's struggling because the real challenges are far more nuanced:

• Most brands simply don't have enough behavioural depth on the majority of their customers to personalise meaningfully - additional predictive work is required.
  
  
• Even when the data exists, it needs to be transformed into something usable - raw purchase history isn't enough.
  
  
• Personalisation tools often fall short in orchestrating diverse, adaptive, and campaign-aware outputs.
  
  
• The internal experience gap means many teams are learning on the job, often making mistakes that set projects back months.
  
  
• And perhaps most critically, personalisation is rarely aligned tightly enough to commercial goals to survive long-term prioritisation battles.

If you're serious about CRM personalisation, it's time to ditch the legacy excuses and start tackling the real work. Because done right, personalisation isn't just a nice-to-have - it's a performance multiplier hiding in plain sight.

Photo by Arantxa Aniorte on Unsplash