From the first build to now

WhatAFind started with a straightforward job: help Canadian shoppers find the products featured in our social videos. Somewhere along the way, it grew into a shopping journal, a home for personal product writing, a mascot-led brand, and a question about who was actually going to run all of it.

The site now centres on product discovery again, with Our Finds and a separate Blog. Getting there involved building things, changing my mind, and removing work that no longer belonged in the same place.

September 16 — getting the first version working

The original brief was a mobile-first catalogue: product images, names, short descriptions, and clearly labelled links to retailers serving Canada. Purchases would happen on the retailer’s website. There was no reason to build a cart or checkout.

I also wanted a way to add, edit, and remove products without coding. The first version included a protected product editor, product detail pages, About, and an affiliate disclosure. The starting content was illustrative, so it needed to be labelled that way. I specifically did not want invented reviews or claims that we had tested things.

The first site worked, but the appearance wasn’t there yet. My reaction was: “It looks nice, but it doesn’t grab my attention.” I wanted something homey, like a cozy shopping blog. That led to another visual direction, then more work on the actual browsing experience.

The cards needed consistent alignment. The spacing was too loose. Carousel padding clipped the cards. There were controls and bits of copy that didn’t need to be there. I kept asking for less wasted vertical space and a more compact layout. Lexend became the body font for easier reading, with Fredoka retained for headings.

September 17–18 — giving the site its own identity

The supplied daisy logo became the basis for the next redesign. The site moved from the early navy treatment into plum, cream, and yellow. I wanted the interface to follow the logo’s visual language, rather than having a logo dropped onto an unrelated template.

The structure expanded too. Our Finds became a fuller catalogue organised into categories, while the Journal created space for product writing and shopping guides. I removed redundant navigation, revised active states, and kept adjusting the cards for smaller screens.

There were already two different reasons to visit: someone might want to read about a product, or they might have just seen it in a video and want the link. Individual find cards gained direct retailer destinations. The catalogue then moved toward compact, image-led tiles organised by room and use.

The first personal review covered the Bob and Brad mini massage gun. About also began explaining the family experience and my product-design perspective. That introduced a distinction that mattered throughout the project: personal experience and researched product information needed to remain identifiable.

The original product-management interface was removed during this phase. Although I had asked for self-service editing at the start, the site’s content was now being maintained through the development workflow. That would become important later.

September 18–20 — the Journal and Atti phase

The Journal became much more than a secondary blog page. It developed a newspaper-style front page, with a lead story, different editorial sections, product imagery, and shopping notes. On September 19, it became the homepage.

The writing grew to include oral care, body care, and other product experiences. Original photos and specific personal notes were added to the body-care story. Product strips and direct shopping actions helped connect the writing to the products being discussed.

Atti became the mascot and narrator. The distinction was deliberate: Atti could introduce the stories and react to them, but the family supplied the actual product experience. The voice needed to be plainspoken. A cat presenting a review did not mean pretending the cat had used the product.

I explored reaction stamps and worked them into the editorial pages. Then I reduced their size. Then I removed them from Our Finds. The mascot had a place, but I didn’t want every part of the site covered in stamps.

The daisy also became part of the carousel controls. I wanted the cards to extend toward the screen edges without putting a conventional scrollbar behind the motif. This took several passes through spacing, image scale, and control placement.

Search and filters were added, then made smaller so they did not overwhelm the Journal opening. Product cutouts were enlarged where they were getting lost. Phone typography and the catalogue controls were revised. A lot of this phase was looking at the actual screen and saying: this is too big, this is too far apart, this picture needs more room.

September 21–22 — making the site usable as an affiliate project

By this point, the work went beyond layout and writing. The site needed contact, disclosure, analytics, search visibility, and reliable product destinations.

The custom domain, search metadata, sitemap, and retailer-click tracking had been introduced. Analytics gained consent controls. A contact form was built with validation and delivery safeguards. The recorded contact test reached the email provider’s delivered state; inbox placement was still a separate check.

The affiliate-readiness review exposed issues that visual polish couldn’t resolve. Image sourcing needed clearer records. Manually collected Amazon price ranges had been added, then were removed following the review. Shopping actions could take readers to the retailer for current price and availability instead.

Affiliate links and disclosure wording were updated together as account setup progressed. Product matching still had to be handled individually. A similar-looking item was not enough to establish that it was the product in the story or video.

Search work also had to be described accurately. Some pages were recorded as indexed, and indexing requests were submitted for others. A request was not the same thing as appearing in search, and none of that established that the site was earning money.

September 22 — separating the work and the responsibility

This was the biggest change in the project.

Who owned the site, who supplied the content, and who would maintain it had become separate questions. I explored a complete handover, then a separate PicknPack/Linktree direction, then returned to transferring WhatAFind. The direction changed more than once.

The personal Journal writing, Atti assets, and associated catalogue entries were saved separately and removed from the active site. The homepage returned to product discovery. The new About copy described the catalogue without presenting every listed item as something we had personally tested.

That content separation was implemented. It did not, by itself, complete a transfer of hosting, domain, email, or affiliate accounts. Those remained separate ownership tasks.

A Videos page and a publishing interface were also built during this transition. I then asked to remove the Videos page and the videos. The next instruction was more specific: let this chat update Our Finds and Blog when new videos appear on Instagram.

That became the publishing direction. Instagram would supply the source material; the website would hold the product entries and useful written explanations.

September 23–24 — rebuilding the Blog around product explanations

The Instagram import mapped 13 product reels to 12 explainers because two clips showed the same phone mount. The personal Aquaphor/Atti carousel was excluded. These were product explainers, with no invented personal testing.

I then went back through how the new Blog looked and worked. I wanted pictures with backgrounds, a layout that was easier to read, a sticky search area, thinner tabs, and less vertical bulk. Redundant headings were removed. The catalogue cards went through another compaction pass as well.

For cards with a guide, the card itself became the route into the guide, while the retailer action stayed available separately. This revised the earlier emphasis on sending people straight from a card to a shop: the supporting explanation needed to be reachable too.

The content workflow became more explicit. Known Instagram post URLs were recorded to prevent duplicate imports. Unidentified products could appear as clearly pending catalogue entries with a source link, without a guessed retailer match. That did not automatically justify writing a full article.

I also asked to complete the technical search improvements and deepen three existing articles before generating more posts. Responsive images, metadata, internal links, and article content were updated. The aim was to improve what was already there before adding more volume.

September 24–26 — resolving links and maintaining the catalogue

The latest recorded releases concentrated on matching products and completing their shopping links.

The GASLAND cooktop was matched to the black model shown in the source rather than a stainless-steel alternative. The garden saw was identified as a corded DEKOPRO pole saw, which meant correcting the earlier unresolved identification and battery discussion. A supplied short link subsequently resolved the blower listing.

The project records publication of the matched SiteStripe links, followed by the supplied short links in version 125 on September 25. The GASLAND article and promotional-image rights remain marked as pending in the source ledger.

Recurring Instagram checks were set up to compare source URLs with the publishing record, update supported content, and stay quiet when nothing changed. The recent check history reports no new publication needed. This is a local Codex workflow that depends on the computer, app, and source access being available.

Where it stands

WhatAFind now has a catalogue-led homepage, room-based browsing, product search, a separate Blog, and retailer links. The personal Journal and Atti phase remains part of its development history, while that material is excluded from the active publishing workflow.

←→

The maintenance question is still practical: how much can run from a checked source, and what still needs a person to identify the product, confirm the image source, or resolve an account issue? The workflow now makes more of those boundaries explicit.

There is a working site and a record of published changes. There is not yet evidence in this development record to claim revenue, improved conversion, or a completed transfer of every account. Those are still separate things to establish.