Forty-two percent of recorded mobile interactions on a standard commerce site are the result of interface unresponsiveness rather than any genuine desire to explore. You likely believe that when your analytics dashboard shows a spike in “events per session,” your users are finally discovering the value you’ve worked so hard to build. The truth is often more violent. Behind those clean, upward-trending lines is a person whose patience has dissolved into a rhythmic, percussive assault on a piece of glass that refuses to acknowledge their presence.
42%
Intent
Percentage of interactions driven by interface latency rather than user exploration.
The Staccato Force of Frustration
You see Sofia standing on a train platform, the damp air of a clinging to her coat, as she tries to tap a “Continue” button that has become momentarily possessed by a heavy JavaScript execution. She taps it once, politely, expecting the immediate feedback of a page transition; she pauses for , her brain registering the lack of a visual response; she taps again, this time with a sharp, staccato force that carries the weight of her morning’s mounting frustrations; she strikes a third time, nearly a punch, just as the main thread finally clears and processes all three inputs in a frantic, overlapping sequence that skips her past the next two screens entirely.
You are watching a user who is now lost, irritated, and likely to abandon the cart, yet the person who opens the analytics software on Monday morning will see three interactions on a high-value button and conclude that the new UI is “compelling.”
The Feedback Loop of False Positives
You are looking at a fundamental failure of translation between human intent and digital measurement. We have built an entire industry on the assumption that a click is a vote of confidence, but a counter is a primitive instrument that cannot tell the difference between a kiss and a slap. When a product is slow, it doesn’t just lose users; it creates a feedback loop of false positives that reward the very sluggishness that is killing the business. If your site takes three seconds to respond to a tap, and the user taps five times in that window, your engagement rate has just quintupled on paper while your brand equity has collapsed in reality.
You understand the lie of the proxy metric once you look at the incentives. A product manager’s career might depend on “increased interaction depth,” a metric so vague it encompasses both the delightful discovery of a hidden feature and the frantic rage-clicking of a user who is trying to figure out why the “Cancel” button isn’t working. We treat data as a neutral record of history, but data is actually a selective memory, choosing to remember the touch but forgetting the context of the touch.
The Pogo-Sticking Fallacy
You can see this play out in every corner of the modern web, where “pogo-sticking”-the act of jumping back and forth between a list and a detail page-is often hailed as high engagement even when it’s caused by the user being unable to find the information they need on the first attempt.
The dashboard records a victory.
The dashboard records a surge.
The dashboard records a user base that is deeply invested in the journey.
The dashboard records everything except the silent scream of the person holding the phone.
The Systemic Rot of the Proxy Goal
You probably recognize this sensation from other areas of life where the measurement has replaced the goal. It is the same systemic rot that affects call centers where “average handle time” is the primary KPI, forcing representatives to hang up on complex problems just to keep their numbers in the green; it is the same logic that leads hospitals to measure “throughput” over patient outcomes, turning the healing process into a logistics exercise; it is the same blindness that permits a classroom to be judged by “participation counts,” rewarding the loudest child rather than the most thoughtful one. In each of these cases, the number wasn’t lying, but it was being asked a question it wasn’t equipped to answer.
Call Centers
Handle Time > Resolution
Hospitals
Throughput > Outcomes
Education
Volume > Insight
It begins with a slight stutter in the rendering engine as a third-party tracking script hogging the main thread prevents the browser from responding to the user’s input; it progresses to a missed frame as the unoptimized CSS recalculations force the layout to thrash; it results in a user tapping a second and third time out of a misplaced sense of duty to a machine that has seemingly frozen; it creates a cluster of events that the backend logs as high-intent engagement; it finally manifests as a colorful bar chart in a boardroom that convinces the stakeholders their most sluggish, bloated feature is actually their most beloved asset. You cannot fix what you cannot accurately describe, and if your description of success is “more taps,” you will inevitably build a product that requires more taps just to function.
You find companies like
who refuse to play this shell game with performance data. They recognize that site speed isn’t a “nice-to-have” marketing checkbox, but a contractual obligation that protects the integrity of user behavior.
When you measure interaction to next paint (INP) at the 75th percentile of real-user data, you aren’t just looking for a technical score; you are looking for the moment the user loses faith in the interface. By committing to specific performance thresholds-largest contentful paint under 2.5 seconds, stable layouts with minimal shifting-they are essentially promising that the “engagement” you see on your screen is actually the engagement you intended to foster.
The Condition of the Locked Door
You know the specific, hollow feeling of being locked out. Yesterday, I stood in a parking lot, staring through the window of my car at the keys sitting on the driver’s seat. I pulled the handle once, then twice, then rattled the door with a frantic, useless energy that did nothing to change the physical reality of the lock. If my car were an app, it would have recorded those three tugs as “revisited interest in the entry sequence.” I was engaged, certainly, but I wasn’t happy, and I wasn’t getting where I needed to go.
This is the condition of the modern web user: a person rattling a locked door and being told by the landlord that the building has never been more popular.
You are the denominator in a calculation that doesn’t care about your time. When developers choose frameworks based on what is trendy rather than what is fast, they are taxing your attention to pay for their resume building. When marketers layer on tracking pixels that choke the browser’s ability to process your clicks, they are burning your battery to watch you fail. This is why the industry needs more than just better tools; it needs a shift in accountability where the vendor’s fee is tied to the actual performance experienced by the user. If the button doesn’t work in under two hundred milliseconds, the developer shouldn’t be congratulated for the “high engagement” that follows.
Demanding Digital Fluidity
You deserve a digital world that responds with the same fluidity as a physical object. If you pick up a glass, it moves; if you turn a page in a book, it flips. The insolence of the digital interface is its ability to ignore you while simultaneously claiming it is listening to you. We must demand a standard of performance that treats the user’s thumb as a sacred input, not a data point to be harvested through frustration. Any measurement that cannot distinguish between struggle and success will eventually be used to justify the struggle, and we have spent far too long justifying the stuttering, lagging, and jumping of our digital experiences.
You can change the question. Instead of asking how many times a button was pressed, ask how long the user had to wait before the world changed in response to their touch. Instead of celebrating the volume of data, interrogate the quality of the experience that produced it. The next time you see a spike in engagement on a slow-loading page, don’t open the champagne. Open the code, find the bottleneck, and apologize to the users who were just trying to get on with their day. The metrics are not the truth; they are just the shadow that the truth casts on the wall of your office.