Customers Left Clues. Did the Business Notice?
Every Click Says Something
Customers rarely submit a report explaining what went wrong. They leave clues instead: a series watched to the end, a video skipped after two seconds, a ride cancelled, or a checkout abandoned.
Each signal looks small on its own, but together they form a behavioural trail that can reveal intent, friction and unmet needs.
The opportunity is not simply to collect these signals. It is to use them twice. First, intelligent automation that can respond in the moment and then product analytics and business intelligence tools that can help teams study patterns across the customer journey and decide what should change next.
Netflix: Automating the Next Best Choice
Netflix shows the first use case clearly. Its recommendation system uses members’ interaction histories and contextual signals to rank titles and personalise rows such as “Continue Watching” and “Today’s Top Picks for You” (Hsiao et al., 2025).
Get Joy C. Chukwukere’s stories in your inbox
Join Medium for free to get updates from this writer.
That is more than reporting. The system interprets behaviour and acts without waiting for a manager. It reduces search friction and makes the experience feel relevant. However, automation does not replace product & business analytics. Without product analytics, experimentation, model monitoring and user research, teams may not recognise when automated recommendations begin to misread user intent.
TikTok: When the Automation Is the Product
TikTok goes further because the recommendation engine is central to the product itself. Its “For You” feed considers signals including watch history, likes, shares, comments and searches (TikTok, 2022). Every interaction updates what the system believes the user wants, and the next recommendation follows almost immediately.
This creates a responsive experience while influencing which creators and topics gain attention.
Bolt: Studying the Wider Pattern
Bolt shows what happens after automation acts: teams still need to study patterns. Using Mixpanel, Bolt analyses funnels, retention, cohorts and behavioural patterns across its mobility products. According to Mixpanel’s case study, Bolt used their platform, and their insights contributed to a 3% reduction in ride cancellations and helped Bolt free about 15% of Android engineering capacity after reducing support for Android 4 (Mixpanel, n.d.).
The dashboard showed where to investigate, but people still had to interpret the pattern and make the decision
Behavioural analytics adds another layer: heatmaps show where users click or stop scrolling, while session recordings reconstruct journeys so teams can see confusion that a conversion percentage may hide (Microsoft, 2026).
The Real Advantage Is Using Data Twice
The strongest businesses build a closed loop: observe, act, measure and learn. Recommendation systems and intelligent automation improve the experience proactively by presenting relevant content, actions or pages. Dashboards, funnels, cohorts, heatmaps and session recordings then reveal whether those interventions worked and what the business should improve next.
Using only automation makes a business responsive but potentially blind. Using only analytics makes it informed but slow. The advantage comes from doing both: act on customer data in real time, then return to the evidence to make better product, operational and strategic decisions.
