Nobody outside your company cares that your website exists. That is the uncomfortable starting point, and every website launch press release that fails has failed because it was written by someone who did not believe it.
This is not an argument against announcing a launch. It is an argument about what makes the announcement land. A new site, on its own, is a maintenance event. Companies replace their websites every two to four years as a matter of course, which means the fact of a redesign carries roughly the same news value as repainting an office. The releases that work carry something else, and the site is how you found out about it.
What follows is the four-test filter that decides whether to send one at all, then the structure that works when the answer is yes.
The Four Tests
Call this the Launch News Test. Four questions, and you need a yes on at least two. One yes means you have an internal memo. Zero means do not send anything and put the budget into the thing you will announce next quarter.
Test one is the capability test. Does the site do something your company could not do before? Not look better, not load faster, do something. A booking system where there was a phone number. A configurator where there was a PDF. A self-serve signup where there was a sales call. New capability is a real change in how customers interact with you, and it gives a reporter a reason to care because it describes a business decision rather than a design decision.
Test two is the data test. Did building it produce a number nobody has published? Teams running a replatform almost always generate genuinely interesting data and almost never use it. How many of your customers were on mobile versus what you assumed. How many support tickets were caused by one broken form. What the old site’s conversion rate was against the new one after thirty days. These numbers are specific, they are yours, and they are the single most reliable way to make a launch release worth a journalist’s time.
Test three is the strategy test. Does the site mark a visible change in direction a reporter could write about without you? Entering a new market, dropping a product line, changing who you sell to, going from enterprise to self-serve. The site is the artifact. The direction is the story.
Test four is the stakeholder test. Is there an audience that needs to be formally told, on the record, at a known time? Franchisees, channel partners, regulators, investors, a distributor network. This is the only test that justifies a wire release on procedural grounds rather than news grounds, and when it applies, the other three stop mattering.
Run the tests honestly. The most common outcome, by a wide margin, is one yes on test four and nothing else, which means you need a partner email rather than a press release.
Two things make the filter work in practice. Run it before the launch date is set rather than the week of, because a capability that would pass test one can often be shipped alongside the site if someone decides it matters three months out. And run it with the person who built the thing, not the person who commissioned it, because the builder knows which changes were substantive and which were cosmetic, and they are usually surprised that anyone wants to know.
The test that gets gamed most often is test two. A team will reach for an analytics number, find nothing interesting, and then manufacture a statistic from a customer survey of eleven people. That is worse than having no number, because a reporter who checks the methodology stops trusting everything else in the release. If the data test produces nothing real, mark it a no and move on.
Write the Release Around the Test You Passed
The structural mistake in nearly every website launch press release is that the lead describes the site. The lead should describe whichever test you passed, and the site should arrive in the second paragraph as the evidence.
Open with the change. If you passed the capability test, the first sentence says what customers can now do that they could not do on Monday. If you passed the data test, the first sentence is the number. If it was the strategy test, the first sentence is the direction, stated plainly enough that someone outside the company understands what shifted.
Then give the site one short paragraph. What it is, what it replaced, when it went live. Resist describing the design. Nobody has ever assigned a story because a release mentioned a cleaner interface or an improved user journey, and the words used to describe design are the exact words that mark a release as having no news in it.
The quote goes third, and it should carry information rather than enthusiasm. A chief executive saying they are excited about the new site adds nothing and costs you credibility. A chief executive saying they moved the whole product behind a self-serve signup because 70% of inbound leads were never going to take a sales call is a quote a reporter can use, because it contains a decision and a reason.
Close with the specifics a journalist needs and nothing else. The URL, the date, what changed on the back end if it is relevant, a named contact with a mobile number. No boilerplate paragraph about your mission. No awards.
One more structural note, on the headline. A headline that says a company has launched a new website is self-cancelling, because the reader learns the entire contents of the release in nine words and has no reason to continue. A headline that names the capability, the number, or the direction gives them a reason. The URL belongs in the body, never in the headline, and the word “redesign” is worth avoiding entirely because it signals maintenance.
Keep the whole thing between 300 and 400 words. Wire services price by word count with aggressive overage, and more importantly, a release that runs long is read as having padded a thin story, which is usually correct.
Decide Where It Goes, Not Just That It Goes
Distribution choice is where launch budgets quietly disappear.
A wire release makes your announcement broadly and simultaneously public. That is valuable when test four applies and a partner network needs an on-the-record moment. It is close to worthless as a visibility play, because wire syndication produces many near-identical copies of one story across many domains, and search engines consolidate those down to roughly one result. The search footprint that looks impressive in week one has mostly evaporated by week six.
Individual placements behave differently. Across the Instant Press catalog of 1,016 publications, 96% are Google-indexed and the mean Domain Authority is 62.3, with 174 titles above 80. Each of those is a separate editorial page on a separate domain with its own authority, ranking on its own. Three placements on indexed publications, each covering the angle from a slightly different direction, hold their positions for years. One wire release covering the same news does not.
Timing inside the day matters less than people think and timing inside the quarter matters more. A launch announced in the same week as a funding round, a product release, or an industry conference competes with itself for attention that was always going to be finite. Spreading three announcements across three months produces three search footprints and three chances at coverage. Bundling them into one week produces one story that mentions all three and two wasted opportunities.
There is also a direct route people skip. If your launch passed the capability or strategy test, two or three reporters who cover your sector would genuinely want to know, and a 90-word email to each of them is free. The failure here is almost always sequencing: teams buy the wire first, publish the news, and then email the reporter about something that is already public and therefore no longer worth their time. Email first, wire second, always.
For a launch that passed only test two, the data test, there is a fourth option worth considering. Publish the numbers yourself, in detail, on the new site, and pitch the analysis rather than the launch. A genuinely interesting dataset about how customers behaved during a replatform is a story that journalists cover, and the site becomes the place the story lives rather than the thing the story is about.
What To Do When You Failed All Four
Most launches fail all four tests, and that is a normal outcome rather than a failure of the project.
The correct response is to not send a website launch press release and to redirect the attention somewhere it compounds. Three things are worth more than a release nobody reads.
The first is making the site itself findable. A redesign is the single most common cause of a traffic collapse, because redirects get missed, URLs change, and metadata gets rewritten by someone who has never seen a search console. Spending the release budget on a post-launch technical audit protects revenue that the release was never going to generate.
The second is building the thing you will announce. Launches are weak news because they are universal. A capability, a dataset, a direction change, or a partnership is specific news, and if you have none of those today, the honest move is to put a date on when you will.
There is a fourth, and it is the one most teams resist because it feels like admitting defeat. Tell your existing customers properly. A clear email to the people who already use the site, explaining what moved and what they will need to do differently, prevents the support spike that follows most launches and is read by more relevant humans than any release will reach. It is not press, and it generates more value than press would have in this specific situation.
The third is the search result itself. A new site does not change what appears when someone searches your company name, and for most buyers that search result is the actual first impression. Indexed editorial placements change it. A press release announcing a redesign does not.
So the summary is short. Four tests, two yeses required, lead with whichever test you passed rather than with the site, email reporters before you pay a wire, and if you failed all four, spend the money on redirects and build something worth announcing next quarter.