Making ACCEL's service catalog easier to discover through HTML links

ACCEL Digital Agency · Updated 2 November 2026

Disclosure: This is a case study of ACCEL's own website. It was prepared with AI assistance using preserved implementation evidence and automated source checks. Independent human editorial review has not been recorded.

ACCEL's catalog contains 255 services and 1,926 package options. This work connected existing product pages through ordinary HTML links while preserving the catalog cards, prices, images and ordering interface.

The starting point

The original catalog included 24 cards in its initial HTML. JavaScript revealed additional services through search, filters and SHOW MORE. A source-link audit found 231 product pages without an incoming HTML link from another page. Sitemap inclusion alone did not provide those internal connections.

Connecting the catalog

The catalog now includes links to all 255 product pages in its initial HTML. The interactive view still starts with 24 cards and reveals more through SHOW MORE; the complete catalog is available when JavaScript is disabled. Existing landing-page buttons were connected to 184 matching catalog products. Later, 71 links from existing FAQ answers completed contextual coverage across all 255 products.

The implementation uses ordinary a elements with href destinations, following Google's guidance on crawlable links. The generated catalog markup is shared by visitors and crawlers.

What the evidence shows

  • 255 product pages are linked from the shop's HTML.
  • No product is an HTML orphan.
  • Every product also has an additional context source.
  • Each product is two link steps from the home page through the shop.

The original implementation was checked with 34 targeted tests and 16 desktop/mobile comparisons. Its production acceptance compared 536 HTML pages with the approved build and checked eight catalog states, including filters, SHOW MORE and the no-JavaScript view.

The loading trade-off

During the original catalog implementation, compressed catalog HTML increased from 19,219 to 44,959 bytes, and DOM elements increased from 1,318 to 5,709. A separate 13,407-byte compressed listing request was removed. The combined transfer increased by 12,333 bytes while one loading dependency was removed. These are historical implementation measurements; they are not current field performance results.

Maintenance check

On 20 October 2026, an offline check of release v76 verified all 1,531 plaintext asset hashes against its immutable manifest before reading 536 HTML documents. The catalog still contained 255 services and 1,926 packages, with no HTML orphan products. All products retained an additional context source, including 71 FAQ connections, and the same two-step distance from the home page.

The separate v76 acceptance checked its 241 changed URLs and 301 approved copy replacements. The graph maintenance check made no additional production requests and did not repeat browser testing.

2 November 2026: source and live-fetch checkpoint

A later source check covered 575 known page URLs. All 561 URLs in the current sitemap were reachable from the home page through ordinary HTML links; none was an HTML orphan. Source checks also confirmed crawl and indexing directives and self-referencing canonical URLs for the 530 pages listed as discovered but not indexed in Google's report dated 23 October. That dated report does not assess the latest website changes.

Three separately authorized live URL Inspection tests checked the home page, App Store promotion guide and Spotify playlist follower package. Google's smartphone Inspection Tool successfully fetched all three and reported that crawling and indexing were allowed. These were test-tool fetches, not evidence of routine Googlebot crawling or inclusion in the index.

A separate Googlebot rule now excludes query-string variants of the catalog root from crawling. Checks of the published package retained access to all 561 sitemap URLs, including the catalog and product pages, and preserved the other crawler policies. Filter behavior for visitors was unchanged. This addresses unnecessary catalog URL variants; it does not establish why Google had not indexed the 530 reported pages.

What this does not establish

Link coverage is an implementation result. These checks do not establish indexing, improved rankings, search traffic, AI citations, qualified leads or sales. Those outcomes require separate observations over comparable periods. Field performance also needs its own measurements.

The practical lesson is to inspect links in delivered HTML alongside the sitemap, test the customer interface, and document the cost and limits of the implementation.

Browse the service catalog · Read service and ordering questions

Find a service

SEO with AI ↗SMM with AI ↗GEO & AI search ↗Business chatbots ↗CRM development ↗Explore every service ↗