The Three Silent Killers of Personalisation: Why Good Strategies Still Fail

Chapter headingChapter subheadingIn this article, Sita Kalluri, Director – Marketing Optimisation, Fitch Group, explores why good personalisation strategies can still fail – and argues that the biggest barriers are often not technological.Contents

Thank you for reading Infographics | Landing Page Template

3-4 November 2026 | Hilton Kensington | London

The Three Silent Killers of Personalisation: Why Good Strategies Still Fail

Speaker Article by Sita Kalluri

Director - Marketing Optimisation

Fitch Group

INTRODUCTION

Personalisation has never been more technologically sophisticated. Brands can capture more data, predict behaviour, and automate experiences across increasingly complex customer journeys.

Yet having the technology to personalise does not guarantee that the experience will feel relevant.

In this article, Sita Kalluri, Director – Marketing Optimisation, Fitch Group, explores why good personalisation strategies can still fail – and argues that the biggest barriers are often not technological.

From organisational silos and unnecessary complexity to optimising before understanding, she identifies three silent killers that can undermine customer relevance long before the algorithm becomes the problem..

Over the years, I have become convinced that personalisation rarely fails because organisations lack technology.

We have more customer data than ever before. We can detect intent, predict behaviour, automate journeys and generate content. Yet, after all these years, we still haven’t perfected personalisation.

Customers still receive irrelevant messages. They still encounter disconnected experiences across channels. And they still wonder how an organisation that appears to know so much about them can sometimes understand so little about what they actually need.

The problem, in my experience, is rarely one spectacular failure. It is usually a collection of smaller mistakes that build over time and quietly undermine what personalisation was supposed to achieve.

I think of them as the Three Silent Killers of Personalisation.

Killer #1: Organisational Antibodies

This is perhaps the most fundamental mistake.

Personalisation is often treated as a marketing capability when, in reality, it depends on the whole organisation.

Customer data may sit with one team, technology with another, journeys somewhere else, while sales, service and relationship teams hold valuable context that may never make its way into the systems driving personalised experiences.

Questions such as who owns the customer journey, who sets priorities, who controls the budget and who defines success can remain unresolved for far too long.

The result is not always visible resistance. More often, it shows up as slow decision-making, competing priorities, duplicated activity and different parts of the organisation acting on different versions of the same customer.

Customers feel that immediately – they do not distinguish between marketing, sales, service, digital channels or relationship teams. They experience one company.

This becomes particularly important in relationship-led environments, where some of the richest customer intelligence may not sit inside the technology stack at all. It can live in conversations, relationship history and the context held by people closest to the customer.

We have become very good at capturing what customers click, and we are often much less good at incorporating what customers tell us. Technology can connect data, but it cannot, by itself, align an organisation.

The lesson is simple: Personalisation is not a marketing project. It is an operating model. Customers do not care which department personalised; they simply experience the company.

Killer #2: More is Not Maturity

There is an understandable temptation in personalisation to equate scale with maturity.

More data, audiences, journeys, rules, content, and automation. The activity can look impressive. New use cases are launched, more signals are introduced, and more interactions become automated; but more is not always better, and sometimes we are simply creating more ways to deliver an experience that still is not particularly relevant.

We add another journey before fixing the friction already created by the journeys in place. We introduce another rule without understanding how it interacts with others.

We optimise individual touchpoints while making the overall customer experience increasingly difficult to navigate.

Eventually, a very simple question becomes surprisingly difficult to answer: Why did this customer receive this particular experience? That should concern us.

Personalisation is supposed to reduce friction and make experiences more relevant. If the machinery behind it becomes so complicated that we struggle to explain the outcome, complexity has started working against us.

The lesson is that activity is not the same as progress, and relevance matters more than volume. The most mature personalisation programme is not necessarily the one doing the most.

Sometimes good personalisation means removing a step, reducing a communication or deciding there is no useful reason to intervene at all.

Killer #3: Optimising Before Understanding

AI has made personalisation significantly more powerful.

We can identify behavioural patterns, detect intent and predict the likelihood of a customer taking a particular action with increasing sophistication. That capability is enormously valuable. But prediction and understanding are not the same thing. A model may tell us that someone is highly likely to buy, enquire, engage or leave. It does not necessarily tell us why.

The why matters.

Customers make decisions because of changing priorities, timing, relationships, organisational pressures, risk appetite and a host of other factors that may never appear cleanly in a dataset.

There is another dimension to this too: Personalisation is supposed to be evidence-led, but internal assumptions can easily shape the experience before we have really understood the customer.

A senior stakeholder believes a proposition should work. A team has invested heavily in a capability and wants to demonstrate its value. A product needs greater adoption. An audience is assumed to be interested because it fits an existing hypothesis.

Slowly, the question changes. Instead of asking what the customer needs, we start asking how personalisation can help deliver the outcome the organisation has already decided it wants.

This is where experimentation becomes so important, not because every interaction needs to become an experiment, but because testing permits organisations to be wrong.

It shifts the conversation from: “We know this will work.” to “We believe this may work. Let’s find out.” That distinction matters.

A surprising amount of personalisation still begins with: “What do we want the customer to do?” Click. Register. Download. Respond. Buy. Attend. Convert.

Of course, businesses need commercial outcomes, and personalisation cannot exist separately from them.

But there is another question that should come first: “What is the customer trying to achieve?” That small shift can completely change the experience we design for the customer.

Instead of asking how to create another touchpoint, we might remove friction. Instead of sending another recommendation, we might surface the one piece of information someone actually needs. Instead of maximising engagement, we might decide there is no reason to engage at that moment.

The lesson is that we should not optimise before we understand.

Machine intelligence can help us identify patterns and possibilities. Human context, experimentation, and customer evidence help us understand what those patterns actually mean.

The purpose of personalisation is not to demonstrate how much we know about somebody. It is to make what happens next more useful.

So, What is the Biggest Mistake?

If I had to reduce all three killers to one underlying problem, it would be this: We have become so focused on our ability to personalise that we sometimes forget to ask whether the customer experiences any value from it.

The next era of personalisation will undoubtedly involve more AI, richer data and better prediction. Those capabilities will continue to advance. But I do not think the organisations that succeed will simply be those with the most sophisticated technology.

They will be the ones that organise around the customer rather than around functions, distinguish meaningful progress from activity, and resist the temptation to optimise before they truly understand.

Ultimately, great personalisation comes down to understanding intent well enough to know when to engage, how to engage and, just as importantly, when not to engage at all.

Personalisation rarely fails because organisations cannot personalise. It fails because they are not organised around the customer; they confuse more with better, and they optimise before they understand.

Speaker Article by Sita Kalluri

Director - Marketing Optimisation

Fitch Group

3-4 November 2026 | Hilton Kensington | London

The future of personalisation will not simply be defined by how much data organisations can collect or how intelligently they can automate. It will depend on how effectively they can turn customer understanding into experiences that are genuinely useful.

As AI, first-party data and real-time decisioning make personalisation increasingly scalable, getting the fundamentals right becomes even more important: breaking down silos, reducing unnecessary complexity, understanding customer intent and measuring value rather than activity.

These are the questions at the heart of the Marketing Personalisation Summit 2026, taking place 3–4 November at Hilton Kensington, London.

Join Sita Kalluri and senior marketing, CRM and eCommerce leaders to explore how organisations can resolve the personalisation paradox – delivering AI-powered relevance at scale without losing the trust, context and human understanding that make experiences meaningful.

DOWNLOAD THE AGENDA

BECOME A SPONSOR