How a 32-person inventory SaaS earned 247 AI citations in 8 weeks

The starting picture#

Acme Inventory (composite name) is a 32-person B2B SaaS selling inventory ops software to DTC e-commerce brands between 10 and 250 employees. Before they came to Visibly, they had a strong organic presence — #3 on Google for “inventory management software,” #1 for two long-tail variants, and roughly 14,000 monthly organic sessions. By every traditional SEO metric they were winning.

But on the question that increasingly mattered for their buyers — “what inventory tools should I use for a DTC business?” — they were invisible. ChatGPT cited two competitors. Perplexity cited three different competitors. Claude cited a Forbes listicle. AI Overviews surfaced a Reddit thread. Across all five surfaces, Acme appeared zero times.

The audit, run in week 0, ranked twelve opportunity gaps. The top one was clear: their cornerstone comparison page (“Acme vs. competitor A vs. competitor B”) was where Google sent buyers, but it was written in 2023, had no schema markup beyond a generic Article tag, opened with a 240-word marketing intro, and had no dateModified. The page was structurally invisible to LLMs even though it was visible to humans.

The plan we wrote (and the one we actually executed)#

The opportunity scoring said: rewrite the cornerstone, retrofit schema across the top 5 pages, publish two new comparison pieces targeting adjacent buyer queries, ship one free tool (an inventory turnover calculator). Total: 4 articles plus 2 retrofits in 8 weeks.

The unusual part is what we didn’t do. The customer had 30 monthly article credits and we used 4. The audit explicitly recommended against adding more — their topical cluster was already dense enough; the leverage was in restructuring, not adding.

The week-4 hold was the most important decision. Citations had started moving by week 3 (ChatGPT was at 18; Claude at 4; Perplexity at 27). The natural impulse is to accelerate — ship more content, double down. The audit said the opposite: don’t add new pieces while the existing rewrites are still being re-crawled, because new content dilutes cluster signals while a model is forming an opinion about which pages on the site are authoritative.

What we saw in the monitor#

Every Friday the Visibly weekly report landed. Here are the actual numbers, week by week:

• Week 1: 0 citations baseline (matches audit).
• Week 2: 3 citations — all from Perplexity, all on the rewritten cornerstone within 36 hours of the schema retrofit going live.
• Week 3: 18 citations cumulative. ChatGPT picked it up; Claude still silent.
• Week 4: 31 citations. Claude finally cited at day 18 from the retrofit. AI Overviews still absent.
• Week 5: 64 citations. Free tool indexed in ChatGPT Search; we hadn’t expected this to fire so fast.
• Week 6: 112 citations. AI Overviews started surfacing the cornerstone. Gemini followed within 48 hours.
• Week 7: 178 citations.
• Week 8: 247 citations across all five surfaces.

The schema that did most of the work#

The schema retrofit on the cornerstone was the single highest-impact change. Here’s the structure we landed on. It’s not exotic — it’s just rigorous.

{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "The best inventory software for DTC brands in 2026",
  "datePublished": "2026-03-04",
  "dateModified": "2026-05-12",
  "author": {
    "@type": "Organization",
    "name": "Acme Inventory"
  },
  "publisher": {
    "@type": "Organization",
    "name": "Acme Inventory",
    "logo": {
      "@type": "ImageObject",
      "url": "https://example.com/logo.svg"
    }
  },
  "mainEntityOfPage": "https://example.com/best-inventory-software"
}

Below it on the same page sits a FAQPage block built from the section’s existing Q&A content (no new copy needed — we just marked up what was already on the page). And above the article wrapper sits a BreadcrumbList. The three together are what made the cornerstone structurally citation-eligible.

A glossary for the rest of this case study#

Cornerstone page

The single page on a domain that earns most of its commercial intent traffic and citations. Identifying yours correctly is the audit’s most consequential output.

In-cluster authority

The strength signal a site earns by having multiple adjacent pages on the same topic that link to each other coherently. Strong cluster authority is why a smaller site can outrank a larger one on the topic both cover.

Citation latency

The time between publishing/updating a page and the first AI citation it earns. Perplexity is typically fastest (hours); Claude is typically slowest (2–3 weeks); ChatGPT Search sits in between (days).

What we’d do differently next time#

Three things. First, audit the customer’s existing content templates for date inheritance bugs before week 1 — we’d have saved three weeks if we’d caught the 2023 date in week 0 rather than week 3. Second, ship the free tool earlier; the tool ended up being the single highest-citation-per-week artifact and the audit had it at week 5 rather than week 2. Third, set explicit expectations with the customer about the week-4 hold — they were anxious about the pause despite the data showing it was working.

How this maps to the Audit + Content product#

For prospects evaluating Visibly, the relevant question is “could my team replicate this?” The honest answer is yes — every step above is replicable. The role Visibly played was:

1. The audit identified the cornerstone, the schema gaps, and the opportunity ranking on day 1.
2. The monitor told us which interventions were moving citations and which weren’t, with weekly granularity across all five surfaces.
3. The content engine produced the four pieces that needed producing, in the format the audit recommended, deployed via the customer’s WordPress adapter without any manual handling.
4. The honest reporting flagged the week-4 hold, the 2023-date bug, and the week-6 inflection in real time, so the customer could react.

If you want to see what the audit would surface for your site, run the free audit. It takes about 10 minutes and you get the diagnosis whether or not you ever buy anything.