Techbanq

Problem · performance

Four seconds before the first rate is visible

The hero image is shipped at desktop resolution to a phone on hotel wifi. By the time a price appears, a share of the people who wanted one have opened the aggregator app instead.

Enterprise delivery credentials

  • ISO

    ISO/IEC 27001

    ISO/IEC 27001 certified

Certifications & registrations

Where the time goes

Five causes, and what each one costs

Ordered by what they typically recover on a hotel site. The first two are almost always the whole problem.

VariableEffect on priceWhy
Images at the wrong sizeUsually the largest single costA 3MB hero scaled by the browser rather than by the pipeline. Responsive sizes and a modern format typically recover more than every other fix combined, and they change nothing a designer cares about.
Marketing tagsGrows quietly over yearsEvery campaign adds a script and none are removed. An audit that lists each tag against the person who asked for it usually retires half of them, and that is a conversation rather than an engineering task.
FontsDelays the first readable textSeveral weights, both scripts, loaded before anything renders. Subsetting and a sensible fallback recover a second or more, and Arabic typography needs this attention more than Latin does.
The booking engine's own weightOutside your codebase, inside your funnelA slow engine cannot be optimised by you, but it can be loaded later, measured separately and raised with the vendor with numbers attached.
Rendering strategyDecides the floorA client-rendered site with content in JavaScript has a floor no amount of tuning gets under. Server rendering or static generation for content pages is the structural fix, and it is worth planning rather than retrofitting.

Ask any supplier quoting this

  • What device and connection did you measure on?
  • What is the page weight budget, and what happens in the pipeline when a change exceeds it?
  • Which of these five causes is the largest on our site, with the measurement?

Measurement

Measured on the studio laptop, experienced on hotel wifi

A site tested on office fibre and a recent phone will pass every check and still be slow for a guest on a mid-range device in a lobby. The measurement that matters is the one taken on the hardware and networks your analytics say your guests are actually using, and that data is already in the reports nobody segments.

Once measured, the fix list is usually short and unglamorous, and it stays fixed only if a budget is enforced in the build. Performance regressions arrive one campaign at a time, and a document that says the site should be fast has never stopped one.

Keeping it

Three ways the gain is given back

  1. 01

    A campaign adds three tags

    Each one small, none removed afterwards. A tag inventory with an owner and a review date is the only thing that has ever worked here.

  2. 02

    A property uploads originals

    The content model has to resize on upload rather than trusting whoever is publishing. Anything else fails within a month of handover.

  3. 03

    A new video on the homepage

    Autoplaying video is the single most expensive thing a hotel homepage can carry. It can be done well — poster image, lazy load, no autoplay on cellular — but it has to be built rather than dropped in.

Our role

What a performance engagement does

  1. 01

    Measure on real conditions

    Mid-range devices and the network profiles your own analytics show, not a synthetic score from an office connection.

  2. 02

    Fix the two that matter

    Images and scripts carry most of the weight on a hotel site; we do those first and measure again before touching anything structural.

  3. 03

    Enforce the budget in the pipeline

    A build that fails when a page exceeds its weight, so the gain survives the next campaign without anybody policing it.

Where the booking engine itself is the slow part, that is a vendor conversation — we give you the measurement to have it with.

Common questions

What buyers ask

Will this hurt how the hotel looks?+

Not if it is done properly. Serving an image at the size it is displayed, in a modern format, is invisible to a viewer and is most of the gain. The trade-offs only start at things like autoplaying video and very large galleries, and those are worth discussing individually rather than removing on principle.

Does site speed affect rankings?+

It is one signal among many and we will not promise a ranking movement from it. The reason to fix it is the funnel: a guest who leaves before a rate renders was not a ranking problem. If rankings improve as well, treat it as a second-order benefit rather than the business case.

How long does the first pass take?+

Measurement in a few days and the image and script work in two to three weeks on most sites. Structural changes — rendering strategy, a new content model — are a project rather than a pass, and it is worth knowing which of the two you are looking at before committing a budget.

A measurement on the devices and connections your guests actually use, with the three fixes that recover the most time named and costed.

Measure it on a real connection