Citation & Source Influence
Citation gap audit: how to find what cites your competitors but not you
A citation gap is a source that AI engines cite for buyer prompts in your category but that does not yet mention your brand. The 5-step audit takes 2 to 4 hours for a fixed prompt set, classifies gaps by source type and recovery feasibility, and produces a prioritized 90-day source strategy.
Updated 2026-05-06
Questions this guide answers
- How do I find which sources AI cites for my competitors?
- What is a citation gap?
- How do I run a citation gap audit?
Direct answer
Most AI answers in a category are built from a small number of repeat-cited sources: a few subreddits, two or three editorial roundups, a handful of review sites, and the occasional newsletter. A citation gap audit asks one question about each of those sources: are they citing your competitors but not you? The audit identifies those sources systematically against a fixed prompt set, classifies each by source layer (Reddit, G2, editorial, newsletter, vendor-controlled, dead-link), and ranks them by realistic recovery feasibility — not by raw citation count.
The mention-not-link distinction is what makes this different from a backlink audit. The cited source rarely links to a competitor; it typically just mentions them in passing. A Reddit thread saying 'we use [competitor] at our company' can drive citation share that 50 backlinks cannot.
The 5-step audit takes 2 to 4 hours for a brand with 30 priority prompts and produces a 90-day source strategy. If your brand is invisible in AI answers despite having strong owned content, the gap is almost always at the cited-sources layer.
Why this differs from traditional SEO competitive analysis
Traditional competitive analysis maps the keywords your competitors rank for that you do not. It works in a Google-search world where the buyer sees a ranked list of links and clicks through.
AI search inverts this. The buyer asks a prompt, the engine cites 3 to 8 sources, and the buyer sees only those cited sources. The competitive question becomes: which sources are cited for prompts that should retrieve us, and how do we earn presence in those sources?
This is not a backlink question. The cited sources are not always linking to your competitor. They often just mention the competitor in their content. A Reddit thread saying 'we use [competitor] at our company' can drive citation share that 50 backlinks cannot. Citation gap auditing is the discipline of finding those mention-or-not gaps and closing them.
The 5-step audit framework
The audit is designed to be reproducible quarter over quarter. Each step has a fixed output that feeds the next.
Step 1. Build the prompt set
Start with 25 to 30 buyer prompts representing your category. The prompt set is your fixed test grid. Audit always uses the same prompt set so changes are comparable. Setup takes about 30 minutes.
- 10 category prompts: best [category] for [persona]
- 8 use-case prompts: [category] for [specific situation]
- 5 comparison prompts: [your brand] vs [competitor]
- 5 alternative prompts: alternatives to [competitor]
- 2 buyer-concern prompts: is [category] worth it for [persona]
Step 2. Run prompts and record cited sources
For each AI engine relevant to your category, typically ChatGPT, Perplexity, Google AI Overviews, Claude, and where applicable Gemini, run all 25 to 30 prompts. Plan 30 to 45 minutes per engine.
For each prompt, record all sources cited in the answer, which sources reference your brand vs which reference competitors, and the specific text snippet the engine extracted from each source. A spreadsheet with columns prompt, engine, cited source, mentions you (Y or N), and mentions which competitor is enough.
Step 3. Tabulate sources by frequency and brand presence
Roll up the spreadsheet by source URL or domain to identify sources cited 5 or more times across your prompt set, which of those high-influence sources mention your brand vs which do not, and the source type (owned, earned editorial, or community).
- A typical mid-market B2B SaaS audit surfaces 15 to 30 high-influence sources.
- Of those, 10 to 20 are gaps that do not mention the brand.
- 5 to 10 are presence: sources that already mention the brand.
Step 4. Classify each gap
Classify each gap by source type, recovery feasibility, and time-to-recover. A G2 review drive is a 1 to 3 month effort with high feasibility. A Wikipedia entry can take 6 to 12+ months and often fails. A Reddit thread sits in the middle: 3 to 9 months of ethical engagement.
Step 5. Prioritize by influence and recovery feasibility
For each gap, calculate a Citation Recovery Index that combines three inputs: how often the source is cited in your prompt set, how realistic recovery is given the source type, and how relevant the source is to your top buyer prompts. A workable scoring formula is cited_frequency + (10 × recovery_feasibility) + (5 × prompt_relevance), each on a 1-to-5 scale where appropriate. Sort by score; the top 5 to 10 gaps form your 90-day source strategy.
Recovery feasibility by source type
Use this matrix as the input to Step 4. Recovery feasibility is the realistic likelihood that ethical effort can earn brand mention or citation in the source within the listed window.
| Source type | Recovery feasibility | Typical time to recover |
|---|---|---|
| Reddit thread | Medium (engage ethically) | 3 to 9 months |
| G2 / Capterra / TrustRadius | High (run a review drive) | 1 to 3 months |
| YouTube reviewer | Medium (outreach plus product seed) | 3 to 6 months |
| Tier-1 editorial (Wirecutter, Bon Appétit) | Low | 6 to 18 months |
| Niche newsletter | High | 1 to 4 months |
| Vertical forum or Discord | Medium | 3 to 9 months |
| Wikipedia | Low (notability standards strict) | 6 to 12+ months |
| Vendor-controlled comparison page | Medium | 1 to 6 months |
| Dead or removed page (cited from cache) | Not recoverable | Skip |
Illustrative scenario: how the audit looks in a developer-tools category
The following is an illustrative scenario, not a real client audit. It shows how the framework typically produces a prioritized fix list when applied to a developer-tools category with established competitors.
Setup: a brand running a 28-prompt set across ChatGPT, Perplexity, and Google AI Overviews would see roughly 80 to 90 prompt-engine combinations and surface a few hundred distinct cited sources. After rolling up by domain, the high-influence sources (cited 5+ times across the prompt set) typically split into a smaller group that already mentions the brand and a larger group that does not — those are the gaps the audit prioritizes.
| Example source type | Source layer | Cited | Feasibility (1-5) | Relevance (1-5) | Recovery Index |
|---|---|---|---|---|---|
| Industry roundup blog post | Editorial | 12 | 5 | 5 | 87 |
| Subreddit comparison thread | 18 | 4 | 5 | 83 | |
| Major review-site category page | Review site | 11 | 5 | 4 | 81 |
| Niche industry publication listicle | Editorial | 8 | 5 | 4 | 78 |
| Vertical newsletter reference | Newsletter | 7 | 5 | 4 | 77 |
| Adjacent subreddit evaluation thread | 9 | 4 | 5 | 74 | |
| Tech community evaluation thread | Community | 6 | 3 | 5 | 61 |
| Competitor founder blog post | Earned (competitor) | 10 | 2 | 5 | 55 |
| Competitor-controlled comparison page | Vendor-controlled | 6 | 2 | 4 | 46 |
| Mainstream press secondary mention | Editorial | 5 | 1 | 5 | 40 |
What the typical pattern looks like
Across the source layers above, the highest-leverage early wins tend to be Reddit communities (medium feasibility, high relevance) and review-site category pages (high feasibility, moderate relevance). Editorial outreach is slower but compounds. Vendor-controlled pages are skipped because the asset belongs to a competitor. Long-form founder-authored Reddit contributions consistently outperform casual replies in producing measurable citation share movement.
Common audit mistakes
Five recurring failure modes that distort the priority list:
- Counting the same source as multiple gaps. A G2 page cited across 11 prompts is one gap, not 11. Roll up by source URL or domain before counting.
- Ignoring source type when prioritizing. Closing one high-priority Reddit gap is worth more than closing five low-priority footer-link gaps.
- Treating low-feasibility gaps as priorities. Wikipedia citations are influential but very hard to earn; do not put 12 months of effort there when you could close 6 G2 gaps in 3 months.
- Auditing once and forgetting. The cited-source landscape shifts; re-audit quarterly.
- Confusing citation gap with content gap. A citation gap is about sources that cite competitors but not you. A content gap is about topics competitors cover that you do not. Both matter; do not conflate them.
From audit to operations
The audit produces three operational outputs that feed an AEO program.
Source strategy backlog
The top 5 to 10 prioritized gaps with recovery actions assigned to owners drive the 90-day source strategy.
Content brief inputs
For each high-priority editorial gap, the audit surfaces the kind of content the source publishes. Use that as input for outreach pitches and proposed contributions.
Citation share tracking
Monthly re-audit produces citation share trend lines. Combined with a recovery score, you can prove which source actions actually moved citation share.
FAQ
How often should I re-run the audit?
Quarterly for most brands. Monthly if your category is fast-moving, where new editorial sources or communities emerge quickly.
What if my prompt set does not have 28 prompts?
Start with 15. The framework works at 15+ but produces noisier signal. Below 10, the audit is not statistically meaningful.
Should I audit my competitors' citation gaps too?
Yes. Understanding which sources cite no one in your category surfaces blue-ocean source opportunities. If a high-influence source mentions only one competitor, you have a clearer path than if 5 competitors all already have presence.
What if a high-priority gap is uncloseable, like a competitor's own blog?
Skip it. Competitor-controlled sources are uncloseable, and effort there is wasted. Focus on sources you can ethically influence.
Can citation gap audits also surface outdated or wrong information about my brand?
Yes, and that should run as a separate audit, covered by AI hallucination risk monitoring. The two audits use overlapping methodology but answer different questions.
Do I need to audit ChatGPT, Perplexity, and Google AI Overviews separately, or can one stand in for all three?
Audit at least 2 engines. Citation patterns differ enough that a single-engine audit misses important gaps. Three engines is ideal.
Related guides
AEO Fundamentals
The Answer Gap Is the New Content Brief
Learn what an AI answer gap is, why it matters for AEO, and how marketing teams can turn weak AI answers into practical content briefs.
Prompt Intelligence
AI Search Prompt Set
A practical guide to building an AI search prompt set across category, comparison, risk, implementation, competitor, and brand-specific prompts.
Risk
AI Hallucination Risk Monitoring
AI hallucination risk monitoring helps brands detect inaccurate, outdated, or unsupported claims in AI-generated answers and turn them into governed correction workflows.
Free AI visibility audit
Find out where your brand is missing, miscited, or misrepresented.
SolCrys maps high-intent prompts to mentions, citations, answer accuracy, and content gaps so your team can prioritize the next pages to ship.