Case Study: From Invisible to #1 | Delilah Agency
Los Angeles + Orange County / Est. 2024
← Case studies

From invisible to #1 in our own backyard.

The exact engine we sell, run on ourselves first: our brand-new website went from not ranking at all — not even for our own name — to #1 on Google for the searches our customers actually type.

The starting point

On September 29, 2026, we ran our baseline check on our own brand-new domain. The honest result: not ranking on a single tracked search — and a search for our own name pointed at an old Carrd page, not this website.

We sell search visibility. So before selling it to anyone, we ran the engine on ourselves — and logged every check, good or bad.

The results

Positions below are dated snapshots from live Google searches. Every one is in our tracking log.

Search a customer would typePositionVerified
instagram repost campaign service Los Angeles#1Oct 29, 2026 — also #1 in checks on Oct 24, 25, and 26
instagram repost service Orange County#1Oct 29, 2026 — also #1 in checks on Oct 24, 25, and 26
Delilah Agency#1Oct 29, 2026
instagram repost campaign#3Oct 29, 2026

How it happened

  • Sep 29, 2026 — Baseline: nothing ranking. Our own name didn't find us.
  • Oct 20 — First organic ranking ever recorded for the domain: #3 on the Los Angeles search.
  • Oct 21 — First #1.
  • Oct 24–26 — The homepage held #1 on both the Los Angeles and Orange County searches across three consecutive daily checks.
  • Oct 29 — #1 on the Los Angeles search, #1 on the Orange County search, #1 on our own name, and #3 on the head term — in the check that set the baseline for our expanded 30-search tracking set.
  • In the same window, the site grew from 27 pages to 58 pages — every new page published by the engine.

The honest footnote: these are dated snapshots, not a claim that a position holds every single day. Young sites move around in Google's results — ours did, and our log records those days too. That's the entire point of tracking: you see exactly where you stand, at every check, in plain English.

How it was done

  • A fixed set of tracked searches — 30 for Delilah today, covering our services, our cities, and the questions customers ask AI assistants — checked on a schedule, with every position logged. No cherry-picking.
  • Articles published on the site daily (the engine runs two a day for Delilah), each written for a tracked search in an answer-first format, so both Google and AI answers can use them.
  • Structured data on every article — BlogPosting and FAQPage schema, so search engines and AI assistants can read the pages properly.
  • Internal links from every article to the pages that sell.
  • A listicle strategy — the "best of" roundups AI assistants cite most are a named target of the work.
  • One telling detail: Delilah's own Google Business Profile verification is still in progress. These rankings were won by the website and the content engine alone — no map listing, no shortcuts.

How we track AI search

There's no honest tool that reads ChatGPT's or Perplexity's answers directly, so we use a labeled proxy method: we track the question-style searches people ask AI assistants ("what is the best instagram repost service…"), check the answer-style results and roundups those assistants draw from, and record whether the business is named among them. We never report a made-up "AI ranking" — the method is on the label.

Same engine, second property

The engine also runs for OTL Music, our music agency. In its baseline check on October 29, 2026, otl-music.com ranked #1 for "white label spotify playlisting for agencies" — the exact B2B search its agency clients type — and #4 for its own name.

← Case studies

Crew photos in. Signed reports out.

A custom automation we built for a working cleaning company: Eclipse Building Solution sends Delilah the raw photos and notes from a job — and gets back a signed, client-ready Cleaning Activity Report the same day.

The setup

Eclipse Building Solution is a commercial cleaning company based in Diamond Bar. Their crews already document their work the way good crews do — photos and site notes from every job. What that documentation needed was packaging: turned into a finished report a client can file, forward, and sign off on.

So we built them the workflow. It's a reporting automation, not a template someone fills in by hand:

  • Photos + notes in — Eclipse's crew sends the raw material from the job: site photos, notes, times, and what was done.
  • The report gets built — the photos and notes are formatted into a complete Cleaning Activity Report under Eclipse's name.
  • Signed, client-ready PDF out — a finished document Eclipse hands straight to its client as proof of the work.
Page 1 of a Cleaning Activity Report produced by Delilah for Eclipse Building Solution — Pomona, work dated October 31, 2026
Page 1 of a real delivered Cleaning Activity Report — Pomona, work dated October 31, 2026. Produced by Delilah from the crew's photos and notes, and delivered the same day.

The first run

The system's first production work was dated October 31, 2026: two reports for two Pomona buildings, built from that day's crew photos and notes and delivered the same day. When Eclipse sent corrections, those were turned around within the hour.

That's the whole point of the automation: the crew keeps doing the job the way they already do it, and the paperwork their clients want simply exists — finished, signed, and on time.

What this is

This is Delilah's Client Reporting product running in production at a real company — and it's the kind of system our Custom Websites & Automations work produces: take the raw material a business already generates, and automate the polished thing its clients actually see.

To be precise about scope: this case study covers the reporting workflow only — the automation Delilah built and runs for Eclipse's job documentation.