Repeated searches, backtracking and unanswered questions can signal that people are working too hard to find what they need. The effort can return to the organization as calls, incomplete applications or missed opportunities.
The journeys I worry about most aren’t always obviously broken. A journey can look fine in analytics while the person using it feels every point of friction.
Imagine someone arriving with a need that crosses several parts of an organization: a prospective student comparing programs and fees, a resident trying to understand which support option applies to their situation, or a business leader looking for expertise across several practice areas. The organization might see a few searches, a long session, and visits to multiple high-value pages. In a dashboard, that can look like engagement.
But for that visitor, it may mean translating internal language, guessing which department owns the answer and piecing together a decision across half a dozen pages. If they still can’t work out what applies, that effort may return to the organization as a call, an incomplete application, a poorly qualified enquiry or a missed opportunity.
A journey can be working exactly as designed and still be the wrong journey for the person using it.
When someone has to guess which department owns the answer, then piece it together across several pages, they’re doing work your website could have saved them. That’s the user tax.
There’s nothing wrong with searching or browsing. The question is whether someone is exploring their options or still trying to get a straight answer.
Across industries, the pattern can look like this:
| Feature | What the visitor has to do | What the organization sees |
|---|---|---|
| 1. |
Rephrase the need using the organization’s terminology |
Several search queries |
| 2. |
Compare information across departments, services or practice areas |
More page views |
| 3. |
Check requirements, costs or options in different places |
A longer session |
| 4. |
Return to an earlier page after following the wrong route |
Backtracking |
| 5. |
Contact the organization when the answer is still unclear |
A new enquiry |
These signals need context. An exit can mean someone found the answer. A long visit can be useful research, and an application may be saved and completed later. Check the pattern against what people were trying to do before calling it a problem. Look for evidence that people got what they needed, and where they still had to do extra work.
Most dashboards describe activity: searches, page views, time, clicks and exits. They rarely tell you whether the person understood what applied or knew what to do next.
The evidence of user tax is often already there, but it is split across search analytics, web analytics, forms, service desks and audience research.
Patterns worth investigating include:
Compare search language with the pages people reach and the questions frontline teams hear. Look for the same needs recurring across search, incomplete forms and service inquiries, then use audience research to understand what’s happening.
You may not be able to follow the same person from a search to a call. Start by checking whether the same questions keep coming up in both places. Connecting an individual’s website activity to a later service interaction requires appropriate tracking or direct research, with privacy safeguards.
The goal isn’t to eliminate searching, browsing or human help. It’s to identify where activity is masking an unresolved need.
The visitor pays in time and effort. The organization pays when an admissions team explains entry requirements that already exist online, a government contact center redirects someone to the right service, or a professional services firm has to qualify a broad enquiry because the website couldn’t connect the problem with the right expertise.
I wouldn’t treat every human conversation as a failure. Some needs are complex, sensitive or simply better handled by a person. The avoidable demand is the contact created because the digital experience didn’t resolve what it reasonably could.
That changes the measurement question. Ask whether people reached a useful answer or next step, whether repeated searching and backtracking fell, and whether the remaining enquiries arrived with better context. A useful outcome may be knowing a service doesn’t apply, narrowing the options, finding the right evidence or reaching the right team.
Pick one question that repeatedly becomes a call, incomplete application or support request. Follow it across search, content, forms and service handoffs, noting where the person has to translate terminology, compare information or start again. Then connect that effort to what happens next: do they find a useful answer, leave, return later or contact the organization?
Following one need gives the team a focused place to start. Before changing the content or route, record how often that need appears in calls or unfinished applications. Use rates with a consistent basis for comparison, such as the share of applications started that remain incomplete after an agreed period.
Make the improvement, then check again over a comparable period. Account for changes in demand and allow enough time for saved applications to be completed. Watch a few people try the task to see whether they can find what applies and take the next step with less effort.
Fewer calls alone won’t tell you whether the answer got better. Use the comparison alongside what you observe, and investigate other possible explanations before attributing a change to the website improvement.
Following one need from question to consequence should make the problem more specific. The source content may be unclear or incomplete. People may struggle to find the answer. Or the next step may be disconnected from the guidance that leads to it.
Those problems need different responses.
Squiz Content Intelligence audits website content for AI readiness and accessibility, helping teams identify gaps, ambiguity, and other issues affecting a priority need. Teams can then review the findings and strengthen the source content.
Conversational Search, a capability of Squiz Funnelback Search, lets users ask in their own words and receive answers grounded in approved website content, with links to the sources.
When the journey needs to continue into a form, service or other action, Squiz DXP provides the governed content, reusable components, forms and integrations behind that next step.
The technology alone doesn’t decide what a good outcome is. Teams still need to choose the authoritative sources, define where more context is required and decide when a person should take over.
The most expensive part of a digital journey may be the hidden work that never appears clearly in a page-view report.
So don’t ask only whether people used the journey. Ask:
"How much of our organization does someone need to understand before our website can help them?"
In the Intelligence Era, people can start with the need in their own words. The organization can respond with trusted content and a useful next step, then learn from what remains unresolved.
Is your website making people work too hard?
Use this 2 minute assessment to identify gaps in your current experience and choose a practical place to start.
Use this 2 minute assessment to identify gaps in your current experience and choose a practical place to start.
