Friction is anything that slows, confuses, or stops a visitor from taking an action of value. It is the moment of hesitation before clicking a button. The form field that asks a question the visitor cannot answer yet. The page that loads slowly enough to lose attention. The layout that hides the next step.
Some friction is obvious. Most is not. The skill is in finding it systematically rather than guessing at it.
Friction is not always bad
Before mapping friction, it is worth noting that not all friction should be removed. A certain amount of friction serves a purpose.
A multi-step qualification form creates friction - but it also filters out unqualified leads. A pricing page that requires visitors to request a quote adds friction - but it also ensures the sales team talks to serious prospects. A detailed terms and conditions page slows the checkout - but it builds trust with cautious buyers.
The goal is not zero friction. The goal is intentional friction - friction that exists because it serves the business or the visitor, not because nobody noticed it was there.
The three sources of friction data
Friction lives at the intersection of three data sources. Each reveals something the others miss. Using all three together produces a substantially better diagnosis than any one alone.
Quantitative data
Analytics tells you where people drop off. It does not tell you why.
Start with a funnel analysis - the sequence of steps from entry to action of value. For an ecommerce site, this might be: landing page, product page, add to cart, checkout start, payment, confirmation. For a SaaS product: homepage, features page, pricing page, signup, onboarding step one, activation.
Map the drop-off at each step. If 1,000 visitors reach the product page and 200 add something to cart, the product-to-cart conversion rate is 20%. If 200 start checkout and 80 complete payment, the checkout completion rate is 40%. These numbers immediately show where the largest losses occur.
The step with the largest absolute drop-off is not always the best place to start. A 5% improvement on a high-volume step may be worth more than a 20% improvement on a low-volume step. The maths matters.
Behavioural data
Session recordings and heatmaps show you what people do on the page. This is where the qualitative picture starts to form.
Watch for patterns, not individual sessions. A single visitor who scrolls erratically and clicks randomly is not informative. Twenty visitors who all pause at the same point on the page, or who all scroll past the same call to action without seeing it, is a signal.
Heatmaps aggregate this data into a visual summary. Scroll maps show where attention drops off. Click maps show where people tap and click - including areas that are not actually interactive. Rage clicks (rapid repeated clicks on the same element) often indicate frustration with something that looks clickable but is not.
Behavioural data answers the question: what are people actually doing? It does not answer why they are doing it.
Qualitative data
Surveys, interviews, and user testing tell you what people think and feel. This is the layer that explains the behaviour you have observed.
On-page surveys can be targeted to specific points of friction. A one-question survey shown to visitors who are about to leave the checkout - "What is stopping you from completing your purchase?" - often reveals friction that neither analytics nor session recordings can show. Unclear shipping costs. Distrust of the payment process. Uncertainty about the return policy. These are invisible in the data but real in the visitor's mind.
User testing puts real people in front of the product and asks them to complete a task. The instructions are simple: "Find a product you like and buy it." The observations are often revealing. Visitors get lost in navigation that seemed clear to the team that built it. They misunderstand labels that seemed obvious. They miss calls to action that the designer placed prominently.
A framework for friction mapping
When we run a friction mapping exercise, we work through five categories. These are not exhaustive, but they cover the majority of friction we encounter in practice.
1. Clarity friction
The visitor does not understand what to do next, what the product does, or what will happen when they take an action. This is the most common type of friction and often the easiest to fix.
Signals: high bounce rate on pages with complex messaging. Visitors clicking on non-interactive elements. Heatmaps showing attention concentrated on the wrong parts of the page. User testing revealing confusion about basic features or navigation.
2. Effort friction
The visitor understands what to do but the task requires more work than they are willing to invest at this point in the journey. Long forms, complex configuration, mandatory account creation before purchase.
Signals: high drop-off rates at specific form fields. Partially completed forms. Cart abandonment after seeing shipping costs or delivery times. Session recordings showing visitors starting a process and then navigating away.
3. Trust friction
The visitor understands the offer and is willing to act, but something makes them hesitate. They are not sure the business is legitimate, the product will work, or their data will be safe.
Signals: visitors spending excessive time on about pages, review pages, or terms and conditions. Survey responses mentioning uncertainty or risk. Low conversion rates on pages where the action involves payment or personal information.
4. Speed friction
The page loads too slowly, the interaction feels sluggish, or the system takes too long to respond. Research consistently shows that each additional second of page load time reduces conversion rates by a measurable percentage.
Signals: high bounce rates on slow-loading pages. Drop-off spikes that correlate with page load time. Core Web Vitals scores below acceptable thresholds. Session recordings showing visitors waiting for pages to render.
5. Relevance friction
The visitor arrived with an expectation - set by an ad, a search result, or a referral - and the page does not match it. The intent is there, but the experience does not serve it.
Signals: high bounce rates from specific traffic sources. Low engagement on landing pages receiving paid traffic. Disconnect between ad copy and page content. Visitors arriving on the wrong page for their intent and failing to navigate to the right one.
Prioritising what to fix
Friction mapping produces a list of problems. The list is usually longer than the resources available to address it. Prioritisation is essential.
We prioritise based on three factors.
- Impact. How many visitors does this friction point affect? What is the potential uplift if it is resolved? A friction point on a page with 50 daily visitors is less urgent than one on a page with 5,000.
- Confidence. How strong is the evidence? Friction identified by all three data sources (quantitative, behavioural, qualitative) is a higher-confidence finding than friction suggested by one source alone.
- Effort. How complex is the fix? A copy change can be tested in a day. A checkout flow redesign takes weeks. The ratio of expected impact to implementation effort determines the order of work.
This is a version of the ICE framework (Impact, Confidence, Ease) used widely in product development. We apply it specifically to friction points and use it to build the experiment roadmap - the prioritised list of experiments that the Experimentation phase will work through.
The ongoing practice
Friction mapping is not a one-time exercise. Products change. Audiences change. Traffic sources change. A friction point that did not exist six months ago can emerge after a redesign, a new feature launch, or a shift in the composition of your visitation.
The practice is to build friction mapping into the regular rhythm of optimisation work. Analytics dashboards flag changes in funnel performance. Session recordings are reviewed on a regular cadence. On-page surveys run continuously to capture emerging concerns.
Friction mapping is diagnostic work. It tells you where the problems are and helps you prioritise which to address first. The treatment - the experiments that test specific changes - comes next.
The businesses that do this well develop an intuition for friction over time. They start to see it before the data confirms it. But the data always confirms it first. Start with the numbers, layer on the behaviour, ask the people. Then act on what you find.