Your traffic is fine. Your conversion rate isn't. Somewhere between landing and leaving, people decide against you — and nothing in your analytics tells you where or why.
52 things teams use Crazy Egg to find and fix
Every use case is a real job solved by Crazy Egg: a drop in conversions, a page that confuses customers, or a form that users abandon.
Jump to a category, search the list, or filter down based on your job title.
- use cases
- 52
- features
- 14
- job functions
- 7
- industries
- 11
Popular searches:
No use cases match your search or filters. Try another search or clear the filters.
Diagnose what's broken — 9 use cases
Something is underperforming and nobody can say why. Start here.
-
-
Conversions fell off a cliff last week and nobody can explain it. A sharp drop almost always means something broke rather than something drifted, but analytics shows you the dip without the cause.
-
People start signing up and don't finish. You can see the gap in the numbers, but a completion rate flattens every distinct reason someone quit into one useless average.
-
Visitors are clicking the same thing over and over. That's someone expecting a result and not getting one — the clearest frustration signal you can measure, and usually a bug hiding in plain sight.
-
People keep clicking images, headings and icons that don't do anything. Every one of those is free product feedback: your visitors telling you what they expected to be interactive.
-
Mobile converts at a fraction of desktop and everyone has quietly accepted that as normal. It usually isn't an intent problem — it's tap targets, content order and a CTA below a fold that only exists on small screens.
-
The ads are working. The page isn't. You're paying full price for clicks that land somewhere which doesn't continue the promise that earned them, and your cost per acquisition keeps climbing.
-
Half your page may as well not exist — you just can't prove which half. Teams keep adding sections and never remove any, so the most-argued-about content usually sits below where most people stopped reading.
-
New client, no history, and a lot of questions you're expected to answer quickly. Guessing at priorities burns the first month of the engagement and the client notices.
Fix a specific page — 8 use cases
The page you already know is the problem. Pick it and go.
-
Your most-visited page is also the one everyone has an opinion about. Homepages accumulate compromises until they say everything and land nothing, and seniority usually wins the argument.
-
Your highest-intent page is usually your least-tested one. Visitors arrive ready to buy and leave over plan confusion, comparison overload, or one unanswered question.
-
Visitors arrive interested and leave without adding to cart. Product pages rarely fail on persuasion — they fail on missing information, and a shopper who leaves to find sizing or shipping detail doesn't come back.
-
The most expensive drop-off you have. Everyone in checkout has already decided to buy, so losing them is pure waste — and the causes are usually mundane: a surprise shipping cost, a forced account, a field that rejects a valid entry.
-
You're paying for every visitor and most leave in seconds. Landing pages get one job and a short window — and they usually lose on message mismatch, an over-long form, or proof nobody scrolls far enough to see.
-
Every extra step is another chance to change their mind. Signup, registration and application flows are where interest turns into commitment — and where most organizations lose the majority of the people who wanted in.
-
People start your form and give up partway through. Forms fail one field at a time — a confusing label, an aggressive validation rule, one question that feels too intrusive — and you never hear a word about it.
-
Test and prove — 6 use cases
You have an idea. Find out whether it actually works before you ship it everywhere.
-
You've never run one and you don't want your first to be a waste. Most first tests fail because they change something too small to matter, or run on too little traffic to conclude anything at all.
-
The first sentence most visitors read is usually the least tested thing on your site. Headlines carry more conversion weight than almost any other element and cost the least to change — yet they get settled in meetings.
-
Small element, outsized effect, endless opinions. CTAs fail for three different reasons — nobody sees it, nobody understands what happens next, or nobody wants what's on offer — and each needs a different fix.
-
Not the prices — how they're presented. Plan count, ordering, naming and highlighting all change which option people choose without changing what you charge, and almost nobody tests them.
-
A full rebuild is the riskiest change you can make and usually the least validated. Redesigns ship on taste and deadline, then get defended after launch when it's expensive to reverse.
-
You made it better. Now you have to show it as a number. Design and copy work gets cut first in budget reviews precisely because its contribution stays invisible.
Understand a segment — 3 use cases
Site-wide averages hide the answer. These break the data apart.
-
Two very different audiences, one page trying to serve both. First-timers need context and returning visitors need to get on with it — a page tuned for either usually underperforms for the other.
-
Same page, very different intent, very different results. Paid visitors arrive mid-promise from an ad; organic visitors arrive mid-question from a search. Treating them identically wastes the expensive one.
-
Phone, tablet and desktop are three different products. Device decides what's visible, what's tappable and how much patience someone has — and blended reporting hides the worst experience you offer.
Ask visitors directly — 6 use cases
Behavior tells you what happened. Asking tells you why.
-
The visitors who leave know exactly why. Nobody ever asks them. Behavioral data shows you where people go, but only the person leaving can tell you what stopped them — and only while it's still true.
-
Ask while they're using the thing, not a week later by email. Emailed surveys reach the people who already like you enough to open your email — never the frustrated visitor about to leave.
-
"Too expensive" is rarely about price. It's usually shorthand for "I don't understand what I get" — and those are opposite problems with opposite fixes, which is why guessing costs you either way.
-
You've read it a hundred times. They're reading it for the first time. Teams lose the ability to see their own copy fresh, and comprehension failures look identical to disinterest in the data.
-
Your roadmap input comes from your loudest existing customers. The people whose unmet needs explain why they haven't bought are on your website right now, and nobody is asking them anything.
-
The moment right after conversion is the most honest one you get. The experience is fresh, the customer is well-disposed, and most sites waste it saying thank you.
Personalize the experience — 6 use cases
One message can't fit every visitor. Serve the right one to the right person.
-
You paid for a promise; the page should keep it. Message mismatch between ad and landing page is one of the most expensive leaks in paid media and one of the cheapest to close — no campaign changes required.
-
A hospital and a startup do not need the same homepage. Horizontal products lose deals by sounding generic, and relevance usually beats any headline test you could run.
-
What works in one market quietly fails in another. Currency, language, regulation and which customers they'll actually recognize all change by geography — full localization is expensive, targeted relevance isn't.
-
Stop introducing yourself to people who already know you. Returning visitors have moved past the explanation stage, and showing them top-of-funnel framing wastes the visit.
-
Named accounts, generic website. ABM programs invest heavily in reaching specific companies and then send them to the same page as everyone else — the site is the weakest link in an otherwise tailored campaign.
-
The right offer at the wrong time is the wrong offer. Site-wide promo banners annoy every visitor they don't apply to and dilute the ones they do.
Report and align — 3 use cases
Getting other people to see what you see, and act on it.
-
The recurring task that eats a day and gets skimmed in a minute. Assembling numbers by hand from several tools is slow and error-prone, and the interpretation — the only part with any value — gets squeezed to the end.
-
A chart proves something changed. A recording shows why. Executives discount analytics they can't interrogate, and thirty seconds of a real person failing to complete a task moves decisions that months of dashboards did not.
-
Your findings are only worth what the client understands of them. Agencies lose billable hours turning raw analysis into narrative — necessary work that no client wants to pay for as a line item.
AI and agentic workflows — 9 use cases
Put your behavioral data to work in the AI tools your team already runs.
-
You don't want a dashboard. You want to know what's wrong. Most people responsible for a website aren't analysts and never will be — so reports that require interpretation skill leave most of their value unclaimed.
-
Your team already standardized on a model. Use it. Dumping a raw export into an AI tool means the model spends its budget parsing your file instead of answering your question — and you re-upload the whole thing for every follow-up.
-
An agent that can only read is a reporting tool. Agentic workflows stall the moment the agent can observe but not act — and most systems it can connect to are read-only sources.
-
Diagnosis is cheap. A redesign takes a team. Knowing a page underperforms doesn't produce a better page, and that gap is where most optimization programs quietly stall.
-
AI spend is the one line item still growing, and it's now being scrutinized on output per dollar. The expensive part was never the model — it's that raw exports make it burn budget parsing your files before it can reason about anything.
-
The questions are simple. The queue is not. When every answer has to be requested from someone else, most questions never get asked at all — and the ones that do arrive after the decision was already made.
-
Your executives will never log into an analytics platform. They already ask Copilot, Claude or Gemini for everything else — so if your behavioral story only lives in a tool nobody senior opens, it isn't shaping a single decision that matters.
-
Your behavioral data is the one dataset your models can't get anywhere else — and it's the one still trapped in a reporting tool while everything else has been centralized.
-
You are not short of data. You are short of conclusions. Most teams collect far more than they ever interpret, and the reports pile up because turning a chart into a decision still takes a person with time and the skill to read it.
Operations and governance — 2 use cases
Keeping a large or complex web estate healthy, compliant and consistent.
-
The riskiest week your website will have all year. Migrations break things that pass QA — tracking, forms, third-party scripts, one browser — and the damage compounds daily until someone happens to notice.
-
You need the insight and you need to pass the review. Regulated teams get blanket bans on session recording that usually come from a misunderstanding of what it actually captures.
Not sure which one you have?
That's the most common starting point. Install on one page and Crazy Egg will tell you where visitors are struggling before you've decided what to look for.
