Costs
Part of Influencer discovery gets cheaper when you qualify in the right order
Influencer discovery strategy before the search: portfolio, adjacency, exclusions
Influencer discovery strategy: deciding the portfolio shape before the search, finding adjacency, writing exclusions first, and sizing the sample to learn from.
Two brands using the same database with the same filters will shortlist the same twenty creators. That is not a flaw in the tools; it is what a shared index does. If your discovery process consists of running a search everyone else can run, your competitive advantage in this channel is your budget, which is a bad place to be.
Discovery strategy is the work you do before the search: deciding the shape of the portfolio, deciding where you are willing to look that other people are not, and deciding what you will refuse. The evaluation of individual creators comes after, and influencer discovery covers that part.
What to take away
- Decide the portfolio shape before you look at anyone, because a shortlist assembled first will bend the strategy to fit it.
- The searchable middle of any database is the most competitive and most expensive part of the market, and it is where a filter-driven process always lands.
- A discovery process with no written exclusions will produce a list nobody can say no to.
Design the portfolio before the search
Three questions settle most of it, and none of them requires looking at a single creator.
How many partners, at what scale?
How long is the commitment?
What has to be shown, and where can it be shown?
Where to look when everyone is looking in the same place
Where to look
What it surfaces
- Standard database filters
- The well-cataloged middle, already being approached by your competitors
- Your own customers and community
- People who already use the product and have an audience, however small
- Comment sections of creators you already rejected
- Adjacent creators the index has not connected to your category
- Adjacent categories
- Audiences that overlap yours without being labeled as yours
- Creators who mention the category without being in it
- Genuine enthusiasm, unpriced by the market
- Inbound pitches you have archived
- People who already want to work with you
The cost of using it
- Standard database filters
- Fast, cheap, crowded, and priced accordingly
- Your own customers and community
- Slow to find, and the audience is often the wrong shape
- Comment sections of creators you already rejected
- Requires reading, which nobody wants to do
- Adjacent categories
- Harder to justify internally, and higher variance
- Creators who mention the category without being in it
- Time-consuming, and many will not want brand work
- Inbound pitches you have archived
- Self-selected, so skewed toward those seeking sponsorship
None of these replaces database search. The point is that the database is one frontier among several, and it is the only one your competitors are guaranteed to be using too. A shortlist built entirely from filters will be technically defensible and strategically identical to everyone else's.
Search frontiers and their costs
Frontier
- Database filters
- Cataloged middle
- Own customers
- Existing users
- Rejected creators' comments
- Adjacent creators
- Adjacent categories
- Overlapping audiences
- Category mentions
- Genuine enthusiasm
- Archived inbound pitches
- Already want you
Surfaces
- Database filters
- Fast, crowded
- Own customers
- Slow, wrong shape
- Rejected creators' comments
- Requires reading
- Adjacent categories
- Harder to justify
- Category mentions
- Time-consuming
- Archived inbound pitches
- Self-selected
Cost
- Database filters
- Own customers
- Rejected creators' comments
- Adjacent categories
- Category mentions
- Archived inbound pitches
Adjacency is where the value usually sits
The instinct is to search inside the category. The problem is that inside the category, the audience is already saturated with messages about the category, and the creators know exactly what they are worth to you.
The adjacent search asks a different question: who else holds the attention of your desired audience, for a reason unrelated to your product? A tool brand can look at people building things, not people reviewing tools.
A financial product can look at communities built around the life event that triggers the need, not around finance.
Two tests keep this honest. Can you write the sentence explaining why this audience would care, without using the word "lifestyle"? And would the creator's audience find the partnership odd? If the partnership needs explaining to the audience, the adjacency is too far.
Write the exclusions down first
A discovery process without exclusions produces a list where every name is arguable, and arguable lists get decided by whoever is most senior in the room.
Write exclusions before candidates
- Categories you will not appear next to
- Competitor work inside a recent window
- Audience characteristics inappropriate for your product
- History of undisclosed paid posts
- Material connection the audience would not expect
A history of undisclosed paid posts is a compliance problem. It also signals how this person will treat your agreement.
A material connection the audience would not expect must be disclosed clearly and prominently. The current FTC guidance on endorsements, influencers and reviews is the reference for what that means now, not what it meant when your last policy was written.
Write these before you have candidates. Exclusions drafted after you have fallen for someone will be drafted around them.
One exclusion deserves its own note. A creator whose past sponsored work is hard to identify as sponsored is a risk to you as well as to themselves. What a clear disclosure looks like in practice is set out in the FTC's disclosures for social media influencers.
Set the sample so the campaign can teach you something
If part of the point is to learn what works, the portfolio has to be capable of producing an answer.
Designing a sample that teaches
- Vary one thing at a time
- Hold two of scale, platform, format, audience constant
- Book enough of each type to survive an outlier
- Decide the learning before you buy
Vary one thing at a time where you can. A set of partners who differ in scale, platform, format and audience simultaneously will produce results you cannot attribute to anything. Holding two of those constant across a group is usually possible and makes the difference readable.
Book enough of each type that a single outlier does not decide the conclusion. One partner in a category is an anecdote. This is a budget constraint as much as a design one, and it is a legitimate reason to run fewer variants than you would like.
Decide the learning you want before you buy, and accept that a portfolio optimized for learning is not the same as one optimized for this quarter's result. Pretending otherwise is how brands end up with two years of campaigns and no accumulated knowledge.
What discovery hands to outreach
The output of good discovery is not a list of names. It is a list of names with a reason attached to each one, in a form the first message can use.
What discovery hands to outreach
- Why this person and not a similar one
- What specifically you want from them
- What you saw in their work
- Rejected names with reasons
- Qualification evidence kept with the name
For each shortlisted partner, record why this person and not a similar one, what specifically you want from them, and what you saw in their work that made the case. That record keeps the outreach message from being generic.
A generic message is the main reason good partners do not reply. Creator outreach covers what the message then has to do with it.
Record the ones you rejected and why. A rejected-for-now list with reasons is worth more in six months than a shortlist is today, because circumstances change and the reasoning does not have to be redone.
Keep the qualification evidence with the name, because it is the same evidence the rates and negotiation conversation will need. Arriving at a price discussion without your own view of what this partner is worth to you means the first number spoken becomes the frame.
Bottom line
Before opening a search tool, decide the portfolio shape, the commitment length, and the format constraint. Check at least one frontier your competitors are not using, and write exclusions before you have favorites.
Then hand outreach a list where every name carries a reason, because that reason is what the first message is made of.
Common questions
Should we use one database or several?
One, used properly, plus at least one non-database frontier. A second subscription mostly buys you the same crowded middle through a different interface.
How large should a shortlist be?
Large enough that you can lose your first choices without restarting, which in practice means several times the number of slots you intend to fill. Outreach reply rates make this arithmetic unavoidable.
Is a smaller audience always better value?
No. Smaller audiences are often more engaged and cheaper per person reached, and they also cost the same amount of coordination time each, which is the constraint that actually bites at volume.
How far can adjacency stretch?
Until the partnership needs justifying to the audience. If the creator has to explain why they are talking about your product, you have gone one step too far.
