What an over-designed website costs you in enquiries
A beautiful website can still make the next step hard. Run these practical checks before blaming demand, seasonality or the people visiting it.
The site opens with a video. The headline waits for an animation. The menu has been reinvented. It looks excellent on the designer’s laptop, and nobody can explain why the enquiries are thin.
The design may not be the cause. But it is easy to test before you spend money changing everything else.
Check one: can the page become useful quickly?
Run the URL through Google PageSpeed Insights and read the mobile result first.
If the report has real-user data, pay attention to it. Lab tests are valuable for finding regressions, but field data describes what actual Chrome users experienced across different devices and connections.
The current Core Web Vitals are:
- Largest Contentful Paint (LCP): when the main content becomes visible. A good result is 2.5 seconds or less.
- Interaction to Next Paint (INP): how quickly the page responds to an interaction. A good result is 200 milliseconds or less.
- Cumulative Layout Shift (CLS): how much the layout unexpectedly moves. A good result is 0.1 or less.
Google evaluates those thresholds at the 75th percentile, separately for mobile and desktop. That matters: one unusually fast test on the office connection does not describe the customer experience.
Do not reduce this to a single score. Open the diagnostics and look for the cause: an oversized hero asset, scripts doing work before the page is usable, fonts delaying text, or elements being inserted above content after it appears.
Check two: can somebody act without learning the interface?
Open the site on your phone, turn off Wi-Fi and try to complete the main customer task.
For a hospitality business that may be checking rooms or sending an enquiry. For a clinic it may be finding the location and calling. For a professional service it may be understanding what the firm does and deciding whether to make contact.
Count:
- how long before you can read something useful;
- how many choices appear before the main action;
- how many taps it takes to reach that action;
- whether a banner, chat widget or animation covers it; and
- whether the page still makes sense when an image or script is late.
The number of taps is not a universal score. One clear, considered step can be better than a short but confusing path. What matters is whether each step answers a question the visitor actually has.
Check three: watch somebody use it
Give the phone to somebody who has not seen the site. Ask them to explain what the business offers, find one important detail and start an enquiry. Then do not help.
Where they pause is evidence. Where they tap is evidence. “I would normally go back to Google here” is evidence.
Five quiet minutes of observation often separates a performance problem, a copy problem and a navigation problem more clearly than a month of opinions about the design.
What over-design actually means
It does not mean animation, unusual layouts or strong art direction are bad. CJAX builds experiential work too.
An effect becomes over-design when the visitor pays for it and receives no help in return. The payment may be waiting, uncertainty, motion they did not ask for, mobile data, a missed tap or a contact button that arrives after they have already left.
A cinematic introduction can be appropriate when atmosphere is part of the product. It still should not hold the headline, hide the next step or make a low-powered phone do work before the page becomes useful.
The same applies to “conversion” additions. A pop-up, chatbot and sticky button can each sound sensible. Together they can create three competing requests before the visitor understands the business.
Ask the builder for constraints, not adjectives
Before approving a design, ask:
- What is the mobile performance budget?
- Which content works before JavaScript runs?
- What happens for reduced motion and a constrained device?
- Which effect can be removed without changing the customer journey?
- How will the enquiry path be tested end to end?
- What real-user performance will be watched after launch?
A good answer is specific and testable. “Optimised” is not a budget. “SEO-friendly” is not a measurement. “Mobile responsive” does not show that the page is useful on a slow phone.
On this site, the public performance standard sits beside automated limits for JavaScript, fonts, accessibility, security headers and Lighthouse. The exact numbers will differ for another business. The useful part is having limits that can stop a beautiful idea from becoming an expensive regression.
Fix the obstruction, not the personality
The answer is rarely to flatten every website into the same template. Keep what makes the business distinctive. Remove what delays understanding or action. Load costly enhancements later. Let the ordinary HTML carry the message. Respect reduced motion. Test on the device and connection the customer actually has.
Then measure the enquiry path itself. A faster homepage does not help if the form fails, the notification goes nowhere or nobody knows who replies.
Related: What a booking site actually costs you and CJAX Labs services.
Sources: Web Vitals on web.dev · Google PageSpeed Insights