Spy on Your Competitors: The Prompt That Maps Every Creator They Pay

Every creator partnership your competitor runs leaves a public trail. The posts are visible. The collaborations are visible. And the people in that trail have already proven two things about themselves: they work with companies in your category, and they say yes at a price a company like yours can pay.

Almost nobody looks. Here is how, and then here is the part almost everyone skips.

The prompt

Replace the variables and run it. The two blocks people are tempted to delete, constraints and quality check, are the two that decide whether you can trust what comes back.

ROLE
Competitive influence analyst with Favikon MCP access.

VARIABLES
{{COMPETITOR}}    = Monday.com     # the brand, not the domain
{{NETWORK}}       = linkedin       # one network per run
{{PERIOD}}        = last 6 months
{{MARKETS}}       = France, Germany
{{CATEGORY}}      = project management software
{{FOLLOWERS_MIN}} = 5000

TASK
1. List creators who posted about {{COMPETITOR}} on {{NETWORK}}
  during {{PERIOD}} in {{MARKETS}}.
2. Separate disclosed paid partnerships from organic mentions.
3. For the top 10, compare each sponsored post to that
  creator's own median post over the same period.
4. Map the gaps: which sub-niches, follower tiers, countries
  and formats exist in {{CATEGORY}} that {{COMPETITOR}} is
  NOT using.

OUTPUT
Table 1 — handle | followers | ER | country | paid or organic |
post vs own median % | verdict
Table 2 — sub-niche | tier | country | format |
{{COMPETITOR}} present | creators available

CONSTRAINTS
- You observe relationships, not contracts. Label every fee as
 an estimate, never as fact.
- No mention does not mean no deal. Gifting and undisclosed
 arrangements will not appear. Say this in the output.
- Do not fill a missing metric. Write "not available".

QUALITY CHECK
Table 1 should hold 10 to 40 rows. Under 5 usually means
{{MARKETS}} and {{CATEGORY}} stacked as AND filters. Drop
{{MARKETS}} first, never {{CATEGORY}}.

Why step three changes what you do with the list

Without it you are copying a competitor's homework without knowing their grade.

Plenty of partnerships underperform the creator's own normal content. When that happens it usually means the brief was wrong or the audience was never a fit, and copying it is copying a mistake with extra steps. The creators whose sponsored posts beat their own baseline are the ones worth the money, and there are always fewer of them than the roster suggests.

This is also the comparison that protects you in a budget conversation. "They worked with fourteen creators" is trivia. "Four of those fourteen outperformed their own median, and here they are" is a shortlist with a reason attached.

The gap analysis nobody runs

The roster tells you what your competitor is doing. The gap tells you where you can win without bidding against them for the same five people.

That is what table 2 is for, and it is the half of this exercise that creates an advantage rather than a copy. A competitor who owns the 100K-plus tier in France has almost certainly left the 5K to 20K tier in Germany untouched, because nobody has the bandwidth to cover everything. Those gaps are cheaper, less contested, and they are where a smaller budget beats a bigger one.

Three follow-ups worth having ready

Run these in the same conversation, in this order.

• Who mentions {{COMPETITOR}} organically with no deal in place? That is the better list.
• Of the creators who beat their own median, which have never worked in {{CATEGORY}} outside {{COMPETITOR}}?
• Estimated price for those creators, sorted by cost per engaged follower rather than by fee.

The first one is the sleeper. Someone who already talks about your category without being paid is the cheapest, warmest partnership available to you, and they will never appear in a roster search because there is no roster entry to find.

Two honest limits

• You see the relationship, not the contract. Who worked together is observable. What was paid and promised is not. Treat every price as an estimate and never quote one back to a creator as if you know it.
• No mention does not mean no deal. Gifting, affiliate arrangements and undisclosed partnerships will not show up. Read the map as a strong signal, not a complete ledger.

Both of those belong in the output, not just in your head. If you hand this analysis to someone else, the caveats need to travel with it.

Where to run it

Favy reads 200M+ analysed posts across 9 networks, so this runs on observed content rather than a manual scroll through a competitor's feed. It is available on every Favikon plan. To run the same prompt from Claude or ChatGPT without switching tools, connect the Favikon MCP.