Rebuilt from the archive in three and a half days

The WYSIWYG Co. is a Toronto coaching and organizational-development practice run by Caroline and Aaron Sugarman. They coach executives and leaders, work with teams, and run assessments, including Kolbe conative coaching. The name stands for “What You See Is What You Get”; the logo is the rabbit-duck drawing.

On the evening of September 30, their website wasn’t there. The address pointed at a server that didn’t answer. Google still had the pages in its index, so searchers found links that went nowhere. We don’t know when or why it failed. The job was to get them a site back.

The challenge

Most rebuilds start with something: a login, a database export, a folder of files. This one had no source code, no access to the old WordPress server and no backups.

What we did have was the Internet Archive’s Wayback Machine, which had saved the site 127 times between 2002 and April 2026. So that became the source.

From those snapshots we recovered:

• The words. The home page, the service pages and the about page.
• All 30 blog posts, verbatim. Not summarized, not rewritten. Their posts, as they wrote them.
• The logo. The best copy in the archive was a 450-pixel image. We traced it into a clean vector file so it stays sharp at any size and can change colour for light and dark modes.

What the archive didn’t have

The Wayback Machine saves what it happens to visit, not what you’d want saved. It had no captures of the site’s uploaded images after 2019. It had no captures of the about page after mid-2021. And the associates page the practice added from 2022 on was never saved at all.

We didn’t guess at the missing bios and photos. We built the associates section, and the confirmed wording and photos have to come from the client. An archive gets you most of the way; the owners fill the gaps only they can fill.

The choices

Keep every address. The old blog posts lived at short addresses right off the main domain. We put each one back at exactly the same address. That meant no redirects were needed for the posts at all, only the routine ones that send “www” and plain “http” to the right place. Every link Google had, and every link anyone shared, works again.

Build it on something they own. The new site runs on Astro and EmDash, on Cloudflare’s edge network, in the client’s own Cloudflare account. If we vanished tomorrow, the site would still be theirs.

Let the owners edit everything. Every word on the site can be changed by the owners, on the page itself. Nobody has to call us to fix a typo.

Move fast, then listen. We researched the outage and drafted three design directions the night we found it. The client chose one on October 1, and the first full build took one session of about 65 minutes. Then came two revision rounds on October 3.

The first round was the client’s: go back to our own copy, not your interpretation of it; drop the newsletter; take dates and bylines off the blog; soften the call to action. We did all of it. The second round put on-page editing everywhere, swapped our accent colour for their actual blue, and redesigned the blog so the first post sits near the top of the page instead of far down it, with all 30 posts sorted into seven topics.

The site went live on wysiwygco.com the night of October 3. Three and a half days from “it’s down” to “it’s back.”

What came out

Lighthouse, Google’s page-quality test, measured it the same night. Performance, accessibility and best practices all scored 100, on both mobile and desktop. SEO scored 92. The main content appears in 0.4 seconds on desktop and 1.4 seconds on mobile, and nothing jumps around while it loads. The sitemap lists 47 pages, and search engines are allowed in.

What we don’t know yet is how quickly Google reconnects the old listings to the new pages. Matching addresses are the best start we can give it.

What this says about owning your site

A website feels permanent. It isn’t. It sits on a server someone pays for, behind a domain someone renews, built by someone who may have moved on.

This practice was lucky in one way: the archive had their words. Plenty of small sites are never archived at all.

The real cost of a site disappearing isn’t the rebuild. It’s the days when clients search for you and find an error page, and the content you can’t get back.

Check this week

Three questions for your own site:

1. Do you have the source? If your developer disappeared, could someone else pick up the site? Do you know where the files and database live?
2. Do you have the content? Your pages, your posts, your images, saved somewhere you control. Not only on the live site.
3. Do you have the domain login? The account where your domain is registered, in your name, with a password you know.

If any answer is “I’m not sure”, that’s the thing to fix before you need it.

You can see this site next to our other builds on the examples page, and how we put them together on how it works.