“We know how to work with user feedback, extract the meaning, and communicate it to the client”

We spoke with Adel Gaisin, Chief Product Officer at Watchers, about how their analytics system was built and can be used, how the company protects personal data and complies with regulations in different markets, and how user insights gathered from community chats can change everything.

— What do you consider the most important metrics to evaluate chat performance on the platform?

Adel Gaisin: The amount of any user interaction is important. Often, when evaluating chat integration, only the messages sent are counted, but that’s not entirely accurate. Most users in any chat are readers. However, those who simply read can also react to messages, send animated emoticons, or perform crucial actions on the platform, such as making a purchase or clicking internal links. If you compare users who join a chat and do nothing with those who enter the chat and like a message, the difference between them is substantial. The likelihood of a user who liked a message returning is 50% higher. Of course, if a user has already sent a message, the probability of their returning is significantly higher. But there is another factor, which is context. For example, if a user sends a message to complain about the platform, that interaction may lead to negative metrics.

— Often, in product or promo materials, Watchers chats grow retention. What is standing behind retention in chats, what exactly does this metric represent?

Adel Gaisin: If you see retention in chat, it means users have found value there. If retention drops to zero, it indicates they don’t find value in the feature, because chat is definitely a feature, not a standalone product. When users return to the chat, it shows that a community is beginning to form. This suggests the client’s core audience is taking shape and starting to explore more of the platform’s content and context. For example, on video platforms with TV series discussions, it’s always interesting to see different opinions, what people thought about an episode, who has done it. The majority often seeks out other perspectives. It’s part of how people decide whether to watch something. And if they’ve already made that decision, the chat feature encourages them to spend more time on the platform and to return when the next episode or series ends. So, a single chat feature can set you apart from competitors, and a community moves the platform to a user’s top choice.

In the gaming industry, we often see a direct link between users finding value in chat and taking action, such as making a purchase or placing a bet. Opening a specific game is frequently triggered by something the user saw in chat, such as a message or shared widget.

You should always view metrics in combination

We regularly observe how core communities form and how the most active users become brand ambassadors. As the client’s platform continues to develop, this audience often turns into internal influencers, people who grow within the platform without requiring any marketing spend. They value this because they’ve gained visibility, even if just within that platform. Still, it feeds the users’ ego, and it benefits the platform. It’s a perfect combination.

— Are there any metrics you find overrated? Or metrics you used and then realized that they are not needed for chats?

Adel Gaisin: I reckon, for us, it’s the session duration. This is a crucial metric for products directly connected to content consumption. It seems like our product fits that description. After all, we influence retention and time spent. But first, I’ll repeat myself: we’re not a product, we’re a feature.

Also, session duration isn’t always aligned with our core metrics. For us, the key metrics are often business-related actions, such as transactions, games started, subscriptions, or memberships. There is a final destination that the communication should guide users towards. We are not providing a casual chat where users simply say hello. They can do that if they wish, but what matters to us is connecting a user’s entry into a chat with the main platform we’re integrated with. That’s why we continuously develop widgets that offer value. It could be a copy widget that allows users to repeat after someone’s transactional action, or widgets that lead users to a new game or a new season of a TV show.

In general, time spent or session duration without meaningful actions isn’t a good sign. If a user spends a lot of time on a VOD platform but doesn't watch anything, it means the platform’s discovery is poorly designed and needs improvement. Therefore, you should always view metrics in combination: how much time was spent on the platform through the lens of business actions.

— Can it be that chats have the opposite effect on a platform? For instance, users criticize games, movies, or the platform? Or is this blocked by moderation?

Adel Gaisin: That certainly depends on a client’s strategy and approach. If a client wants the cleanest, most positive chats where all users are satisfied with everything, that can be achieved through moderation. On the other hand, who wants a chat where everyone just admires everything around them? If users send negative feedback about the platform in a chat, the platform’s managers should take notice. We also help platforms manage that feedback. If a particular complaint in a chat appears repeatedly, we recommend that the platform address the issue, respond to the user, and win back their loyalty. Thus, it’s still a relationship between the platform and its audience; we are just a tool, like a phone or a messenger, that users employ to express their experience.

— Do chats influence user authorisation on platforms?

Adel Gaisin: Data on login conversion is always collected by the platform. Unfortunately, I can’t share any specific figures here. But we can definitely observe and statistically confirm that when a user sees dynamic content on a page, especially in a community, discussions, or comments, it grabs attention and encourages guests to log in or register and join the convo. This happens regardless of whether the discussion is positive or negative. Everyone wants to be heard.

That’s why we try to create thematic chats in unauthorised zones. On sports platforms, it can be a discussion around the game, or at the very least, something like a chat lobby. When comparing pages with and without a chat, it’s clearer that users are more likely to log in on pages where a chat is present.
 

— Okay, let’s talk about behavioural insights analysis. How do you analyse them, and how can they help clients?

Adel Gaisin: We can distinguish behavioural insights into two kinds: technical problems, and user experience that isn’t related to technical aspects. For the first kind, we’ve built a system that enables interaction between users and the platform. This helps the platform identify technical issues early and act on feedback. Most often, platforms without community tools rely on support tickets to detect problems. But the truth is, not all users reach out to support. In fact, if a user contacts support, it usually means they’ve already encountered a significant issue, and solving it takes time.

When a user mentions a problem in a chat, there’s a chance the community helps them find a quick and simple solution, meaning there wasn’t really an issue after all. In this case, we reduce the load on support. In other cases, the issue is real, but because it’s raised in the chat, it gets noticed and addressed before a support ticket is even created.

The second kind of behavioural insight comes from users discussing things when nothing is technically wrong. By analysing this, we can determine what users are noticing and enjoying, and that information can influence the development or content strategy of our client. For example, sports chats might reveal that the most engaged users prefer discussing small university tournaments over UEFA leagues. This helps clients discover new ways to market, promote, or develop their offerings. They can see what users are paying attention to. For instance, a user might watch a film and never return, and the platform might wonder whether it’s due to pricing or poor content discovery. If the user types in a chat after watching that they couldn’t find any other similar thrillers, then it’s obvious that the issue isn’t the price or the discovery, but the content itself. In that case, the solution is simply to license more of this specific type of content. We can also analyse more nuanced segments and topics.

—Do clients often draw meaningful conclusions from these insights? For example, do they change their UX or start licensing different content?

Adel Gaisin: We have one quite large client who has completely changed their development strategy based on chat insights. They’ve just changed everything. Initially, they integrated chats as an experiment, began working with user feedback, and then redesigned their product. Now, they focus on community as the core feature of their platform, not just another product gimmick. This has become the direction the whole company follows.

We do have clients like that, but not all of them require such a significant shift. For that kind of pivot, several factors need to align: the technical opportunities, of course; a large user base, and a good fit. We’re not the right match for all types of products and audiences.

 — What segment is not relevant?

Adel Gaisin: It depends on the audience. Suppose a particular audience doesn’t want to form a community. I don’t believe there are completely irrelevant industries, but if I'd to name one, I’d probably pick HR tech. I reckon an HR platform isn’t a place where people genuinely want or need to have discussions, whether it’s about job opportunities or career trends. As I’ve said, we aim to influence business metrics. But I don’t see a strong connection between “I discussed this job” and “I received an interview invitation.” Filling a job position is a long process. It comes with many additional challenges, such as NDAs, and when someone is in the middle of that process, it’s not a time for community discussions.

We’re most effective when the consumption cycle is fast and clear: you discuss a product, then buy it; you talk about a film, then watch it; you discuss a game, then play it. It’s a kind of loop. The cycle can be longer but consistent like joining a membership to discuss daily workouts. HR is a bit different.

— Can we imagine the situation: I integrated a chat with the platform, it works, metrics are growing but just chat metrics. My main business metrics are falling down because users are just sitting and talking there. Retention is stunning but my platform is not about this.

Adel Gaisin: Look, we have a relevant example. A client integrated a chat, and the audience started using it actively. Retention there is now just incredible. Retention, of course, always depends on the domain and industry, but we can use 20% as a benchmark. If your retention is at 20%, you have a good, solid product. If your retention is around 40%, that product rocks, it’s genuinely strong. The client I mentioned earlier, who transformed their entire product into a social one, has around a 40% retention rate. However, the company I’ve just mentioned has a 60% retention rate in chat. These users, they’re like a family. So, what should this client do? They need to bring the main features into the chat. You have a channel where your users genuinely want to spend their time, and that channel is built directly into your product. Don’t waste it. We’re always ready to meet our clients’ needs and to improve chat functionality if we see the potential.

— What tools or frameworks do you use when working with data?

Adel Gaisin: First, we mark all actions to analyse user behaviour. There’s nothing magical here; it’s all clear and straightforward. Then, we define a set of key metrics that we track on a constant basis. So, we analyse the number of sessions, specifically how many sessions a user needs to perform a business action; actually, many metrics are involved. And all the tools we use are our own; we built them ourselves. Because we work with many different clients, and each of them uses different internal tools, we developed a universal solution. First, we provide a part of the data via our back office. This helps clients who don’t have the resources to build analytics on their side. They can just open the dashboard and see real-time updates quickly. Second, we’ve created a more or less universal system that sends events to the client, so they can analyse everything in their own analytics environment. It doesn’t matter which exact tool the client uses; it could be Mixpanel, Amplitude, or some built-in tools. We simply deliver the events, and the client can do with them as they need.

— Do you use AI somehow in your analytics?

Adel Gaisin: We use AI when analysing behavioural insights, that’s the semantic analysis we discussed earlier. It’s necessary to understand what’s happening in the chat and to identify user behavioural trends. Semantic analysis is a part of our analytics system.
 

— How do these insights correlate with the trend of gathering first-party data?

Adel Gaisin: First-party data is, quite literally, communication. So, yes, this still falls under the semantic analysis of behavioural insights; it’s about observing the trends that users are discussing. And, frankly speaking, our chats provide a channel that enables the collection of this first-party data. Let me take a step back and say that users often lie. If we directly ask someone about their opinion of our product, there’s a high chance we’ll encounter a little white lie. In short, some of the information we receive from users isn’t entirely accurate at first, simply because they want to be polite or present a better version of themselves. But in an anonymous community, in a friendly environment, users share honest feedback. They discuss what truly matters to them.

We often see messages like, “Oh, great that you’ve added a chat! Сould you also add this and that?” We know how to work with this kind of feedback, how to extract the meaning, and how to communicate it to the client. It’s also incredibly helpful when users have a space to express negative feelings about the platform. Because if they don’t do it in a chat, they might post a public review. And what might be a minor and easily resolved issue can become a permanent, negative comment. If I have access to a chat, I’ll just mention the problem there, rather than leaving a review on AddAdvice or elsewhere. And when the issue gets resolved within the chat, my loyalty to the platform increases.

— How is personal data protected in this system you’ve built?

Adel Gaisin: Personal data is not part of our DNA, as we’ve always aimed to collect as little personal information from clients as possible. In fact, we don’t really need any personal data, we just need a user ID to create an event and an account on our side. User nicknames are not unique, and there’s no link between user IDs, nicknames, and users’ real names. We also don’t allow phone numbers, emails, or any other personal information to be sent or used in any way in a chat. Therefore, the primary assurance that personal data is protected is that we don’t collect any personal data at all.

— Watchers is a global company, available in many countries. Do you use different protocols to operate in all of them?

Adel Gaisin: Data processing depends on the location where the client operates. We’re subject to GDPR in the EU. However, there are additional regulations in different European countries, for example, in Spain, we’re required to retain all information for seven years after the end of our relationship with a client. In the US, the regulations differ, as they do in Latin America and Asia. In any case, we discuss everything directly with the client, and their legal department approves our framework. The core of our system is simple: we decrypt and analyse data, store it if the regulator requires it, and don’t store it if there are no such requirements. And of course, we carry out research for every country we enter and follow all existing regulations.

***

Please fill out the form below if you'd like to integrate community-based live chat or other social tools into your website or app. We will help you find the best solution for your engagement strategy.