Your site is fast on office fibre. Your customer is on a platform at Union Square with one bar and a train coming.
Website speed optimisation in New York costs $2,000 to $12,000 as a scoped engagement. What makes this market different is the testing condition: a large share of New York browsing happens on a phone with degraded connectivity, so a site that scores well on a desktop connection can still be unusable where your customers actually are.
Core Web Vitals on your own URLs, from field data where it exists and lab data where it does not. You get both numbers and the list of changes each gain is attributable to. A claim that the site is faster is not a deliverable.
Throttled mobile, not office fibre. New York browsing happens underground, on platforms and between buildings. A page that renders in one second at a desk and eight seconds on a degraded connection is an eight-second page.
Modern formats, one correctly sized file per breakpoint rather than a large one scaled down, lazy loading below the fold and a CDN in front. On most sites this single workstream produces the majority of the gain.
Chat widgets, heat maps, two analytics tools, a review widget, three pixels. Each was added by someone who is no longer there. We measure what each costs in load time and hand you the list before removing anything.
It affects search modestly and abandonment considerably. The commercial argument is almost always the second one, and it is measurable within weeks rather than months.
Sites regress. A performance budget in the build, plus monitoring, is what stops the next marketing request quietly undoing six weeks of work.
| Project | Range (USD) | Timeline | What you get |
|---|---|---|---|
| Audit only | $1,000 to $2,500 | 3 to 7 days | Field and lab measurement, prioritised findings, effort estimates |
| Standard optimisation | $2,000 to $6,000 | 2 to 4 weeks | Images, scripts, render blocking, caching, before and after numbers |
| Deep optimisation | $6,000 to $12,000 | 4 to 8 weeks | Above plus server response, font strategy, code splitting, third party rework |
| Ecommerce performance | $4,000 to $12,000 | 3 to 8 weeks | Product and collection templates, app or script audit, checkout path |
| Monitoring | $150 to $700 per month | Ongoing | Field data tracking, regression alerts, budget enforcement |
If an audit shows the problem is the platform rather than the pages, we will tell you that instead of selling you eight weeks of tuning.
| Cause | What it costs |
|---|---|
| Full-size images served to phones | Most of the delay on most sites |
| Six third party scripts on every page | Seconds nobody has audited in years |
| Testing only on desktop and wifi | A page that fails where your customers are |
| Web fonts blocking the first render | Blank text while the connection struggles |
| No caching or CDN | Every visitor paying the full server round trip |
| No performance budget after launch | Six weeks of work undone by one campaign tag |
| Cause | What it costs |
|---|---|
| Full-size images served to phones | Most of the delay on most sites |
| Six third party scripts on every page | Seconds nobody has audited in years |
| Testing only on desktop and wifi | A page that fails where your customers are |
| Web fonts blocking the first render | Blank text while the connection struggles |
| No caching or CDN | Every visitor paying the full server round trip |
| No performance budget after launch | Six weeks of work undone by one campaign tag |
| The question | A good answer | Walk away if |
|---|---|---|
| What will you measure? | Core Web Vitals, field and lab, on our URLs | A Lighthouse score |
| What connection will you test on? | Throttled mobile as well as desktop | Whatever the tool defaults to |
| What is attributable to each change? | A list, change by change | The site is much faster now |
| What stops it regressing? | A budget and monitoring, quoted | We can look again later |
Ask what they will measure before you ask what it costs. A supplier who only quotes a Lighthouse score is optimising a number rather than an experience.
An audit runs $1,000 to $2,500. Standard optimisation runs $2,000 to $6,000. Deep work including server response and code splitting runs $6,000 to $12,000. Monitoring afterwards runs $150 to $700 a month.
Because that is where your customers are. A large share of New York browsing happens underground, on platforms and between buildings. A page that renders in a second at a desk and eight seconds on a degraded connection is an eight-second page.
Images, on most sites by a wide margin. Modern formats, one correctly sized file per breakpoint rather than a large one scaled down, lazy loading below the fold and a CDN in front.
Modestly. It affects abandonment considerably more, and that is usually the stronger commercial argument. The conversion effect is also measurable in weeks rather than the months a ranking change takes.
We measure what each one costs in load time and give you the list before removing anything. Most sites carry several nobody can account for, added by people who have since left.
Yes, on your own URLs, with each gain attributed to a specific change. If we cannot show you the difference, you should not pay for it.
A performance budget in the build plus monitoring that alerts on regression. Without those, the next campaign tag quietly undoes several weeks of work and nobody notices for a quarter.
Then we will say so rather than selling eight weeks of tuning. Some builds cannot be made fast without replacing something underneath, and that is a different conversation with a different number.
Send us the URL and we will measure it on a throttled mobile connection and tell you what the first three fixes are.