At Balkan eCommerce Summit 2026, Mike Korba, Co-Founder of User.com and part of Positive User, delivered a practical session on how eCommerce businesses can manage customer experience at scale through CRM, marketing automation, personalization, and AI.
The presentation focused on one of the biggest challenges for modern marketers: companies collect huge amounts of customer data, but often struggle to turn that data into meaningful actions, better customer experiences, and higher revenue.
The key message was clear:
Data alone has no value unless it helps businesses understand customers and personalize their experience.
The Problem: Too Much Data, Too Little Clarity
Mike opened the session with a very familiar situation for eCommerce managers.
A CEO asks a simple question:
“How many customers did our last campaign bring us?”
But for marketers, the answer is rarely simple.
The data may be spread across:
- CRM systems
- eCommerce platforms
- Google Ads
- Google Analytics
- Email tools
- Loyalty programs
- Spreadsheets
- Dashboards
Each system may show different numbers, different customer names, and different versions of the same user.
The result is a lot of manual work:
exporting data, joining tables, cleaning lists, comparing dashboards, and trying to understand what actually happened.
Mike challenged the common industry phrase “data is the new oil.”
According to him, most companies are not operating like large oil companies with powerful refineries. Instead, they are more like someone standing in the desert with a rusty shovel, knowing there is something valuable underneath but not knowing how to extract it.
From Data Collection to Customer Experience
Positive User works by gathering customer data from different sources, including:
- Websites
- eCommerce engines
- Loyalty programs
- CRM systems
- Customer behavior data
- Transaction history
Based on this data, the platform enables brands to personalize customer experiences across multiple channels, such as:
- Email
- SMS
- WhatsApp
- Web push notifications
- Pop-ups
- Loyalty app messages
- Abandoned cart flows
The goal is not only to collect customer data, but to use it in a way that improves communication, timing, relevance, and conversion.
The Personalization Maturity Curve
A central part of the presentation was the concept of the personalization maturity curve.
Mike explained that the more mature a business becomes in personalization, the greater the potential revenue impact.
The growth is not linear. As personalization becomes more advanced, the return can grow much faster.
Level 1: Mass Campaigns
The first level is the traditional mass campaign.
This is when a business sends the same message to the entire newsletter database, often at a fixed time, such as Wednesday at 10:00 AM.
These campaigns are simple and familiar, but they are not deeply personalized.
Typical performance may include:
- Around 15% open rate
- Around 2% CTR
They can still work, but they treat all customers the same.
Level 2: Dynamic Fields
The next step is adding basic personalization, such as:
“Hi, first name.”
This was once considered a major innovation in email marketing, and it still improves results because the message feels more personal.
According to Mike, dynamic subject lines and personalized fields can improve:
- Open rates up to around 55%
- CTR up to around 4%
However, this is still a relatively simple form of personalization.
Level 3: Triggered Campaigns
The next level is triggered communication.
Instead of sending campaigns only according to the company’s calendar, businesses respond to customer behavior.
Examples include:
- Abandoned cart emails
- Welcome emails after registration
- Follow-up messages after webinar attendance
- Product interest reminders
- Post-visit campaigns
This allows brands to react to specific actions rather than sending generic newsletters.
Typical results can improve to:
- Around 30% open rate
- Around 10% CTR
But Mike emphasized that this still reacts mainly to past behavior.
Level 4: Behavioral Segmentation
A more advanced level is behavioral segmentation.
Instead of segmenting customers only by demographic profiles, businesses can now use real behavioral data.
For example, a brand can create a segment of customers who:
- Visited the website recently
- Previously bought premium products
- Have not purchased in the last 30 days
- Have a lifetime value above €500
This type of customer should not receive the same abandoned cart message as a first-time visitor.
When behavioral segmentation is added to campaigns, results can improve further, with:
- Open rates around 40%
- CTR around 12%
The main benefit is that the communication becomes much more relevant to the customer’s real behavior and value.
Level 5: Omnichannel Personalization
The next step is omnichannel communication.
Mike emphasized that customers do not think in channel silos.
Email, SMS, WhatsApp, app notifications, and push messages are all received on the same device. For the customer, they are simply messages from the brand.
A strong omnichannel flow might look like this:
- One hour after cart abandonment, the customer receives an email
- If they return to the website but do not buy, they receive a push or SMS message
- If they still do not react, they receive a personalized offer in the loyalty app
Each channel has a role and supports the others.
This is where many advanced eCommerce companies currently operate, but Mike argued that AI can take personalization even further.
AI in Marketing Automation Is Not Completely New
Mike explained that many AI concepts in marketing automation have existed for years, even if they were not always called “AI.”
Examples include:
- Product recommendations
- Predictive analytics
- Churn prediction
- Propensity-to-buy models
- Sentiment analysis
- Chatbots
- Automated segmentation
Product recommendations, for example, are based on patterns in customer behavior. If users who buy products X and Y also often buy product Z, the system can recommend product Z to similar customers.
The difference today is that modern AI and large language models make these processes much more accessible, faster, and easier to use.
AI as the “Refinery” for Customer Data
One of the strongest metaphors in the presentation was that marketing automation platforms are like shovels and pickaxes, but AI can become the refinery.
Businesses already have valuable data, but the challenge is extracting insights from it.
AI can help by:
- Cleaning data
- Enriching customer profiles
- Finding patterns
- Creating segments
- Answering business questions
- Generating campaign ideas
- Building workflows
- Supporting personalization at scale
Instead of requiring large teams of data engineers and analysts, AI can make customer data more usable for marketing teams.
UMA: AI Assistant for Marketing Automation
Mike also introduced UMA, an AI assistant built into the Positive User ecosystem.
UMA is designed to help marketers interact with CRM and customer data using natural language.
Instead of manually filtering dashboards, exporting files, or writing queries, a marketer could ask questions such as:
“How many customers do I have from Bulgaria?”
or
“Can you split them by city?”
The assistant can then interpret the CRM data model and provide answers directly.
The future roadmap includes capabilities such as:
- Generating emails based on brand voice
- Creating pop-ups
- Building workflows
- Suggesting customer segments
- Automating campaign setup
- Supporting more advanced CRM actions
The idea is to make marketing automation easier to use and more accessible for teams that do not have technical expertise.
Privacy, Data & AI
During the Q&A, Mike also addressed privacy and data management.
He explained that Positive User stores data on its own servers and currently uses Gemini through Google Vertex, hosted in Amsterdam within the European Union.
The company is building UMA with an LLM-agnostic architecture, meaning that future models can be replaced or adapted, including the possibility of using their own models.
The discussion also touched on a key point:
AI can help clean and enrich data, but centralized data remains critical.
To make AI truly useful, companies still need to collect customer information from multiple systems into one connected customer data platform.
Key Takeaways from the Session
The presentation delivered several important insights for eCommerce and marketing teams:
- Data is only valuable when it leads to action
- Many companies collect customer data but struggle to extract useful insights
- Personalization maturity directly impacts revenue potential
- Mass campaigns still work, but advanced personalization performs much better
- Behavioral segmentation creates more relevant customer communication
- Omnichannel journeys should work together rather than operate as separate silos
- AI can make CRM and marketing automation easier, faster, and more scalable
- The future of marketing automation will be more conversational, automated, and insight-driven
Perhaps the strongest message from the session was that the best marketing tools are the ones that work quietly in the background, helping teams deliver the right message to the right customer at the right time.
AI does not replace marketing strategy, but it can help teams finally use the customer data they already have.
