Rebuilding a small-business digital foundation in 36 hours

The Problem

A site that barely worked

I reached out to the owner, described what I was seeing, and talked about the benefit of making some improvements. He was interested but there was no immediate timeline, just a qualified lead for me.

The owner wanted to meet the same day, which we did.

Making the case to rebuild

Side by side, the difference between the original site and 1 hour concept was obvious. The concept showed what the same business could look like with responsive layouts, clear calls to action, readable content, and photos of his amazing work doing more of the selling.

Once he could see the two experiences together, the conversation immediately shifted away from repairing to replacing.

Priorities

This was going to be a time-driven project. The owner needed the site up. There was no time for customer research. The site, poorly done even 15 years ago, had zero analytics to get insights from.

I worked from three sources:

• A 90-minute meeting with the owner to understand his needs and how he saw his customers
• My own perspective as his prospective customer
• Experience designing and building numerous consumer-facing experiences

I had enough to make a set of low-risk decisions and move quickly.

1. Mobile-first

I had no analytics, but this was a consumer-facing local business so I expected mobile to account for the largest share of traffic. I treated it as the primary experience from the start.

2. Protect search ranking

The old site had almost no meaningful SEO, but it already ranked well for important local searches, despite ample competition.

The rebuild could not put that visibility at risk.

I kept the search foundation intact and filled in what the old site was missing:

• Preserved the simple information architecture
• Converted products, prices and descriptions from graphics into actual text and semantic content
• Added SEO-friendly page slugs and plugged in permanent redirects
• Added metadata for SEOs, sharing, LLMs and schema.org structured data
• Added a sitemap, crawler guidance

3. Showcase the product

The owner had years of strong masonry work sitting on his phone, but none of it was visible on the site.

I gathered the owner’s photos from his phone, texts and Facebook, removed weaker and repetitive shots, and ultimately built the gallery with 75+ examples of completed work.

4. Table everything else

Shipping the new site quickly also required deciding what did not need to be in the first release. Deferred items could be worked after MVP was launched.

• Per the owner, DIY parts rarely sold, so I deferred that page.
• I also deferred the standalone Contact page. In the new design, corrected phone and email were already prominent above the fold and repeated throughout the site. Customers could reach the business without it.
• The gallery didn’t need to be perfect. We can cleanup images, tweak photo order, etc, later.
• We tossed around many other ideas like a lead form on the website (instead of email), a FAQ page, integrated social feed, etc. All cut in favor of progress over perfection.

Building

Designed to stay simple and maintainable

I would be taking over website hosting, so I worked the environment setup and domain setup in parallel with building out the site locally so infrastructure woulds not become a blocker at launch time.

For the site itself, I used WordPress and deliberately avoided file-based customizations.

Using a stock vanilla template and putting all the custom work inside the DB meant WordPress upgrades would not overwrite files and breaks would be highly unlikely. That choice quickly paid off, as shortly after launch WordPress released a major new version. The upgrade for me was a single button click but nothing breaking.

Building in Wordpress also leaves the owner or his designate a standard admin interface to maintain the content solo if they desire.

The layout and colors for the site stayed intentionally straightforward: Masonry-inspired colors, brick-toned calls to action, and a simple block-based layout.

The site’s visual design is not trying to be a showpiece. It is meant to be clean, credible, maintainable and appropriate to the business.

AI as an accelerator

AI materially increased how quickly I could work, but it did not make the decisions.

• Claude Design helped me generate and compare early visual directions.
• Claude Code accelerated page builds and repetitive styling changes across the site.
• Nano Banana 2 helped me clean up about 30 gallery images after the MVP launched, removing people, tools, standing water and other distractions. I never altered the mailbox itself.
• AI tools also helped with image processing, alt text and QA across links, accessibility and SEO.

I still reviewed and corrected the output. Early design passes added unnecessary UI and copy, and some automated image processing was too aggressive.

Results

Launch and follow-up

The core site went live in roughly 36 hours.

During the following week I added the lower-priority pages, refined the gallery, optimized images, reviewed alt text, tightened structured data and finished the remaining polish.

The initial release was intentionally focused on the parts that mattered most. The rest could follow without delaying launch.

What the data shows

The old site had no analytics, so there is no credible before-and-after comparison.

Within 60 days of launch, 78.9% of traffic was mobile, supporting the decision to treat mobile as the primary experience.

Visitors were also spending about three minutes on the Gallery page, suggesting they were taking time to browse the work.

The site now has a baseline for measuring what customers actually use.