You prospect agencies, restaurants, practices, e-commerce businesses. You open your 300-row spreadsheet and send the same message to everyone: "Hi, I offer [service], want to connect?" Out of 300 sends, you get three replies — two of which tell you to stop. The problem isn't your offer. The problem is that you have nothing specific to say about the prospect you're reaching out to. You're cold-contacting at random, and it shows from the very first line.
Now imagine opening your message with a concrete observation about their website: "I loaded your homepage on my phone — the main image took nearly six seconds to appear. I looked into why, and here's what I found." You're no longer one of a hundred cold callers. You're someone who took the time to look, and who shows up with an actionable finding. A prospect's website performance is one of the rare signals that is simultaneously public, measurable, and immediately relevant — especially if you sell dev work, redesigns, SEO, or optimization services.
Mobile performance score of a typical prospect — the kind of number that opens a conversation
The website speed audit feature on outsend.xyz, free in alpha, runs an entire list of sites through the Google PageSpeed Insights API and returns a performance score and Core Web Vitals per row. You turn a "website" column into a "qualification level" column — and instantly spot the prospects you actually have something to say to.
What a website speed audit measures
A speed audit isn't just "the site is slow or fast." Google has standardized three user experience metrics, the Core Web Vitals, that each measure something different: rendering speed, responsiveness, and visual stability. The thresholds below are those published by Google on web.dev (web.dev, Core Web Vitals), measured at the 75th percentile of page loads.
| Metric | What it measures | "Good" threshold |
|---|---|---|
| LCP Largest Contentful Paint |
The time before the largest visible element (often the main image or heading) is displayed. The perception of "it loads fast." | < 2.5 s |
| INP Interaction to Next Paint |
The delay between a user action (click, tap, keystroke) and the page's visual response. The perception of "it responds." | < 200 ms |
| CLS Cumulative Layout Shift |
The amount of layout shifts during loading — the button that jumps just as you're about to click it. The perception of "it's stable." | < 0.1 |
These three metrics are combined into a performance score out of 100, calculated by Lighthouse, Google's auditing engine. That's the number shown on the gauge at the top of the page. A site scoring 90+ is fast; a site below 50 has a real problem that its visitors feel — and potentially the prospects of your prospect's customers too.
How outsend audits an entire list of sites
Auditing one site by hand is doable: you open PageSpeed Insights, paste the URL, read the report. The problem starts at row 50 of your file. Nobody is going to paste 300 URLs one by one, wait for each report, and manually copy the score. That's exactly the bottleneck the outsend speed audit feature removes.
You provide a list of sites (one per line, or directly the "website" column from a file you've already extracted). For each URL, outsend queries the Google PageSpeed Insights v5 API and retrieves the Lighthouse performance score along with the Core Web Vitals (LCP, INP, CLS) as lab data. The output enriches your list with several new columns: score out of 100, LCP, CLS, INP, and a human-readable status (good / needs improvement / poor). All of it is exportable as CSV, JSON, or XLSX.
In practice, you start from a raw file and end up with a list sortable by score. If you don't yet have your prospects, you can extract a list of local businesses from Google Maps with their websites, then run the speed audit immediately after — that's the advantage of having scraping and qualification in the same tool rather than across three separate subscriptions. The speed audit chains naturally with other qualification signals: tech stack detection tells you what the site is built with (a bloated WordPress loaded with plugins often explains a poor score), and dead-checking for broken URLs first filters out sites that no longer respond at all.
To see how this fits into the bigger picture, browse all OutSend features.
Try outsend for free
All-in-one, built for European B2B. Free alpha access by application.
Request free alpha accessFrom score to opening line: factual observation, not a lecture
Here's the tricky part — and the one place where a speed audit can backfire. Showing up and saying "your site is slow" doesn't make you more credible. It makes you annoying. Nobody likes having a flaw pointed out by a stranger who wants to sell them something. The score isn't a weapon; it's conversation material.
The difference comes down to how you frame it. Compare these two openers, based on the exact same number:
- Avoid — "Your site is too slow, it's costing you customers. I can fix that." That's a judgment followed by a hard sell. The prospect shuts down immediately.
- Prefer — "I measured the load time of your homepage on mobile: the main image takes about 5 seconds to appear, whereas Google considers anything under 2.5 seconds a good experience. Often it comes down to image format — happy to share what I saw if you're curious." That's a sourced observation, a clue, and a door the prospect can open or leave closed.
The second version works because it's verifiable (the prospect can check for themselves), specific (their page, their number, not a generic claim), and non-aggressive (you're sharing an observation, not calling out a fault). You come across as someone who actually looked — not a salesperson running a script. It's also an excellent way to qualify: a prospect whose site already scores 95 doesn't need you on this topic — save your energy for the ones you can genuinely help.
This logic fits into a broader workflow: you can plug the speed audit into a no-code prospecting pipeline that extracts, qualifies by score, then prepares outreach — as long as you've first checked your own domain's deliverability, because a perfect opening line that lands in spam is worthless.
The limits: a low score isn't a guaranteed need
Be honest with yourself about what the score does and doesn't tell you. A slow site isn't automatically a warm prospect. Three nuances matter.
The score measures a page, not a business. A homepage at 38/100 might belong to a company that doesn't care about its website at all because all its revenue comes from word of mouth or a marketplace. The technical need exists; the willingness to invest in it may not.
Lab data is not field data. The API-based audit returns a lab measurement (a simulated load under standard conditions). That's perfect for comparing sites at scale, but the real experience of actual visitors may vary depending on their devices and connections. For a deep diagnosis on a specific prospect, real user experience data gives a more nuanced picture.
Prospect context trumps the number. A slightly slow page on a site rebuilt six months ago is a dead end; the same slowness on a site that's visibly been neglected for three years, in a sector where local competitors have polished websites, is a much better opening. The score points you in a direction; your judgment makes the call.
In short: the speed audit is a prioritization signal, not a verdict. It tells you where to look first — not who is going to sign.
FAQ — Auditing a prospect's website speed
What are the recommended Core Web Vitals thresholds?
According to Google (web.dev), at the 75th percentile of page loads: LCP under 2.5 s, INP under 200 ms, CLS under 0.1. Above these thresholds, the experience is rated "needs improvement" or "poor." These three metrics are summarized into a Lighthouse performance score out of 100.
What's the difference between the score out of 100 and the Core Web Vitals?
The score out of 100 is a composite grade calculated by Lighthouse from several lab measurements. The Core Web Vitals (LCP, INP, CLS) are three specific metrics that break down the experience: rendering speed, responsiveness, stability. The score gives you the overview for sorting a list; the Vitals explain why a site scores well or poorly.
Is it legal to audit a prospect's site without their consent?
Yes. A speed audit loads a public web page and measures its response time — exactly like any visitor or browser would. No private data, no server access, no user cookies are required. It's the observation of public resources, with no specific GDPR implications for the measurement itself.
How many sites can be audited in alpha?
No strict cap in alpha. The rate of calls to the PageSpeed Insights API is regulated to stay within Google's service limits, so a large list is processed over time rather than all at once. Results remain exportable as CSV, JSON, or XLSX.
Does the audit slow down or disrupt the prospect's site?
No. The measurement goes through the Google PageSpeed Insights API, which performs a standard page load — equivalent to a single visitor. It's not a load test: no heavy traffic is sent to the target server.
Is the speed score enough to qualify a prospect?
No, and you shouldn't present it as such. The score is one prioritization signal among many: it tells you which sites deserve a closer look, but the actual need depends on the prospect's context (industry, reliance on their website, budget). Combined with other signals — tech stack, sector, how recently the site was updated — it becomes a genuine sorting criterion; on its own, it's an indicator, not a certainty.