Personalization in E-commerce: Balance Between Scale and Experience

This article is a summary of the panel discussion held during E-Commerce Growth Meetup #10, an event bringing together founders and executives from leading digital businesses to share practical perspectives on growth.

The full panel recording is available on YouTube:

The panel featured voices from different corners of the e-commerce ecosystem – travel, subscription, therapy services and analytics, showing how personalization plays out in very different models. Moderated by Ewa Wysocka (CEO of Tribe47), the conversation included Agata Szulc (CEO, Travelist), Paweł Chrzan (Co-Founder, WellBe), Piotr Wawrysiuk (CEO, PsiBufet) and Arkadiusz Wiśniewski (CEO & Co-Founder, Sublime Analytics Platform).

What Personalization Really Means

The discussion began with an attempt to pin down the definition. Personalization can mean many things depending on the business model. For marketplaces, it is often about using data to assist customers in their journey, offering relevant deals or content that feels natural. For subscription products, personalization is built directly into the service, from meal plans to curated experiences. And from the analytics perspective, personalization is not about tactics at all, but about the ability to segment, test, and prove what actually works.

What united these perspectives was a shift from persuasion to assistance. The aim is not to pressure the customer into buying, but to guide them in a way that feels supportive.

“I would say that the best and most effective personalization is when users don’t feel that they are tracked and being forced to do things, but that they are assisted and we are there to help them and not to push them to do things.” – Agata Szulc

Why More Is Not Always Better

A key insight from the panel was that personalization should not be treated as a linear progression. The industry often assumes that mass communication is primitive, segmentation is better and hyper-personalization is the ultimate goal. In reality, each of these approaches has its place. Mass campaigns are still essential when launching new products or creating awareness, and generic communication is often the most efficient way to scale. At the same time, going too far into personalization can create friction for customers or prove impossible to scale operationally.

There were vivid examples of businesses trying to personalize every detail, only to discover that it slowed them down. Hyper-personalized flows requiring customers to share extensive information boosted conversion for a small group, but the majority preferred a faster, simpler path. Similarly, attempts to tailor products individually, down to the gram or the feature, often broke the scalability of the model.

The Costs and Trade-offs

Another thread running through the conversation was cost. Personalization is not just about clever algorithms – it requires infrastructure, data collection and constant optimization. Businesses need to weigh the financial and operational effort against the uplift it delivers. The panelists agreed that many personalization initiatives fail to produce statistically significant results, while simpler, proven methods such as RFM segmentation often drive consistent value.

This creates a paradox: customers say they want personalization and those who engage with personalized journeys often spend more, but the majority do not want to go through the effort. Businesses must therefore design for both: a quick, generic path for the many and a personalized journey for the few who are willing.

Product vs. Marketing

The panel also explored whether personalization belongs more to the product or to marketing. For subscription services or highly individualized offerings, personalization is embedded in the product itself — it is the value proposition. For marketplaces or broader e-commerce platforms, personalization functions more as a marketing lever, adjusting inventory, messaging, and communication.

New technologies are blurring these boundaries. Reverse ETL tools now allow data to flow out of analytics warehouses and into customer-facing systems, enabling real-time personalization across touchpoints. But tools alone do not create value. What matters is whether personalization creates experiences that feel authentic and meaningful, rather than manipulative.

“There are new tools that are called reverse ETL. Getting the data out of the data warehouse to all the tools that can provide personalization, like marketing tools, communication. From this perspective, I see the analytics industry is trying to close the loop and catch up with the business requirements.” – Arkadiusz Wiśniewski

Measuring Personalization Correctly

Perhaps the strongest consensus was around measurement. Personalization cannot be judged by intuition. Lifetime Value is the ultimate measure, but it requires time and scale. In the short term, businesses can look at conversion rates and retention, but always in comparison to a control group. Without such benchmarks, seasonality or external factors can be misread as success.

Another overlooked dimension is customer satisfaction. Automating support with AI or chatbots might reduce costs, but if customers feel trapped, the long-term damage can be significant. Personalization should not come at the expense of human connection, especially in industries where trust and emotion matter.

Practical Lessons for Businesses

The panel concluded with advice for founders and executives thinking about personalization. The most important lesson: don’t over-engineer too early. Before building sophisticated systems, focus on product and customer fit. Personalization can be layered over time, as the business grows and the data pool expands. When it is implemented, it must be tested with rigor, avoiding assumptions and vanity metrics. And above all, personalization should be about the customer’s perception — it should feel like help, not surveillance.

“Don’t try to overengineer at the early beginning. The bigger the business gets, probably the more data you’re going to be having, the more personalized messaging you’re able to provide. If you’re trying to do superbly narrow segments at the very tiny business is probably not the way to go.” – Piotr Wawrysiuk

Conclusion

The EGM10 panel challenged the hype around personalization and replaced it with a more balanced perspective. Personalization works when it supports, not when it pressures. It creates value when it is measured and tested, not assumed. And it drives growth only when it is balanced with broader marketing that builds awareness and scale. The businesses that succeed will not be those that personalize the most, but those that know when to personalize, when to communicate broadly and when to step back and let the product speak for itself.

Watch the full video here!