Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Influencer Marketing › Influencer Strategy

Hiring and Evaluating an Influencer Agency

How to hire and judge an agency that runs creator work: what it does, what stays with the brand, questions, warning signs, reports to test and how to leave.
Guideintermediate20 min readUpdated

What an influencer agency is hired to do

An influencer agency is a business paid by the brand to run part or all of its creator work. The site's guide to creator agents and influencer agencies separates it from a creator's agent, who is paid to serve the creator, and sets out when an agency repays its fee. This page starts after that decision. The brand has chosen to buy the work in, and now has to decide what to buy, from whom, and how it will know the work was done well. A freelancer or consultant hired to run the same work needs the same written scope, as the agency selection guide says of a freelance specialist; as general practice, the questions, the reports and the exit below apply to that hire as well.

The work falls into jobs that can be bought separately. The agents guide's responsibility table already covers vetting, negotiation, contracts, briefs, review, disclosure and reporting, so this table adds the seeding, amplification and comments work an agency can be given, each set beside what stays with the brand.

JobThe agency's partThe brand's part
SeedingAsk creators before sending, put the gifting note in the parcel and keep the send recordThe product, its cost and the decision to move a creator from gift to paid deal
Paid amplificationAsk creators for the platform authorisation early, keep each licence and authorisation date on the record and propose posts from the organic evidenceThe ad account, the budget, the decision to run an ad and the licence end date as a hard stop
Comments and incidentsRead the threads if the scope says so, work from the answer pack and flag anything that meets the controversy routineService complaints, the decision to pause or part, and what the brand itself says in public

These rows extend the agents guide's responsibility table. That table settles accountability for each task; start from it, add the rows this table adds and put names against them before any proposal is requested. The RACI matrix explains why a row needs exactly one accountable role. The brand can hand over every row of both tables or a few of them, and that choice shapes the scope.

The brief an agency works from

The agency selection guide holds that most agency work succeeds or fails on the brief, and the strategy hub already describes one. A one-page creator strategy states the role of creators, the objective and its KPI, the audience, the tier and the number, the selection standard, the terms, the budget shape and the review date, and the hub calls it the brief for whoever runs the campaigns. A brand without one can write it before approaching anyone, starting from the influencer strategy hub. Four items are worth stating in the request in their own right.

  • The ceiling. The highest sum the brand can pay for the scope it has defined and still be glad it did, set from its own numbers as the negotiation guide explains.
  • The label standard. The wording and placement the brand will require, taken from the contracts and disclosure guide, so every proposal answers the same requirement.
  • The records the brand will hold. The table in the next section, copied into the request.
  • How results will be judged. One primary KPI and the source of its data, defined before the campaign, as the KPIs and tracking links guide asks.

Send each agency the same brief and the same facts about the business, and score the replies against criteria fixed beforehand, as the agency selection guide asks for any appointment. The agency request for proposal template gathers the shared brief, the questions and the scoring criteria in one place, and the agency scope of work template supplies the fields to fill once an agency is chosen.

What stays in the brand's name

The agents guide lists what the brand should hold on to even when an agency runs everything. This table builds on that list by asking, for each record, where it lives from the first day and what a proposal should show about keeping it, and adds the ad account, the cost record and the send record and request log.

RecordWhere it livesWhat a proposal should showGuide
Creator list and vetting notesThe brand's own files, built as the work goes onHow each note, with its checks and dates, reaches the brand when it is written and not at the endThe vetting section of the creator tiers and selection guide
Signed agreementsCopies in the brand's folderA sample agreement with the licence, the correction right and the disclosure clause visibleThe contracts guide's clause table
Rights and authorisation recordOne line per creator in the brand's asset libraryWho updates it, and who checks it before any ad launchesThe asset library tutorial and the amplification guide
Links, codes and resultsThe brand's own analytics and shop platform, built to the brand's naming conventionHow links are built, tested and registered, and who owns any short linkThe naming conventions and UTMs guide and the discount code request SOP
Creator insightsStored by the brand, in the form agreed with every creatorWhich measures every creator is asked for, in the same formatThe KPIs guide's section on platform insights
Ad account and accessThe brand's own ad account, with the agency holding the access it needs for the workWhich accounts the agency will need and through what accessThe agency selection guide's section on scope and contracts
Cost recordPer creator and per element, in the brand's finance recordsHow fees, product, licences, agent commission and the agency's own fee will appear as separate linesThe cost table in the ROI guide
Send record and request logThe seeding send record and the log of requests creators sendHow both will be kept in files the brand can openThe seeding guide and the requests guide

As general practice, two of these work as gates and not as weights: the accounts the tracking runs in, and the brand's name as holder of the rights, or a clear route by which they pass to it. A proposal that cannot meet them leaves the comparison however well it scores elsewhere, and the weighted scoring matrix keeps must-haves out of the weights for that reason.

Scoping a first round before a programme

The agents guide observes that an agency can carry a first programme while the brand finds out what it would need to run the work itself. As general practice, buy that as a bounded round with a question attached: whether the agency's vetting holds up, whether its agreements and labels pass the brand's own checks, and whether its report can be reconciled with the brand's data. The statement of work for the round should carry the lines below beside the generic ones in the agency selection guide's section on scope.

Which kind of programme. A few larger creators, some with agents, run differently from a programme of many small creators, which the micro-creator tutorial runs from one standard offer and one roster sheet, and both differ from a seeding programme. Name the one being bought. As general practice, where a roster sheet is used, the brand holds the sheet even if the agency keeps it up to date.

Who contracts, and who pays. In whose name the creator agreements are made, whether the agency pays creators and how each invoice reaches the brand's records, raised against an order, as the purchase order tutorial describes.

Review. A named reviewer on the brand's side and an agreed reply time, as the draft review tutorial asks, with the brand deciding every must-fix and one reply going to the creator.

Disclosure checks. The scope says which disclosure checks the agency makes and which the brand keeps; the brand reads some live posts itself. As the agents guide records, the FTC's endorsement guidance allows responsibility to land on the brand and on any agency, not only the creator, and in the UK the ASA applies its advertising codes to creator posts that the brand has paid for and has control over. Read all of this as direction only; the brand's legal adviser decides what the rules require.

Requests. Creators who write to the brand go to its single owner and into one log, as the guide to answering creator requests sets out; as general practice, the agency passes on any that reach it.

What the brand owes. Product that arrives on time, approved claims, a prompt review and payment on the agreed dates. The contracts guide records that paying creators late is a frequent cause of their refusing to work with the brand again, and an agency cannot deliver what depends on inputs the brand has withheld.

Questions for every shortlisted agency

Put the same written questions to each agency on the shortlist and lay the replies out together. As general practice, prefer the reply that names who does each job and what that person needs from the brand.

Representation and income. Do you represent any creator you would propose, or take payment from any creator or platform you recommend as well as from us? The agents guide asks for the question to be put directly and the answer recorded. If the answer is yes, how will you tell us for each name?

Finding and vetting. Which checks, including brand-safety vetting, do you run on each creator, in what order, and what do you do when one cannot be run? Will you rely on the creator's own insights screenshots or on estimates from a discovery tool, whose data the creator selection guide's FAQ calls modelled? What does a vetting note look like, who sees it and when? Show one for a creator you decided against, with the reason.

Contracts and rights. Whose paper will creators sign, and who holds the usage rights? How are the licence period, the markets, paid use and the platform authorisation written down, and who tracks the expiry? Does the agreement carry a correction right and a disclosure clause that reaches every post in the partnership? How narrow will you keep exclusivity? What will you do when a creator's agent sends their own paper? The contracts guide asks that any use of AI tools to edit, translate or extend licensed footage be written into the agreement, and the AI rules guide for agencies covers what else to agree.

Disclosure. Where does the label wording appear in your briefs, and does it cover every material connection, gifted product and commission links as well as paid fees? What do you check in a draft, on the live post and in the replies under it? Who asks for a correction, how quickly, and where is it logged? Which of these will you report to us, and which will we check ourselves?

Briefing and review. How do you separate must-fix comments from suggestions? Who reviews at your end, how many people comment, and how do comments reach the creator? What do you do when a draft reads like an advert?

Paid amplification. Who asks creators for the platform authorisation, how early, and where does its expiry go? How will you choose which posts to back, and how will you test them against each other? What happens on the day a licence ends?

Tracking and reporting. Which accounts will links and codes live in? How are they built, tested on a phone and inside the in-app browser, and registered? Will the report keep link orders, code redemptions and indirect signals on separate lines, carry a cost for each creator that includes every cost, and say what the methods cannot see?

Comments and incidents. Who reads the threads under live posts, and against which answer pack? If a creator becomes the story, who on your side can pause posts and ads, and do they hold the access to do it? Are you named in our controversy routine?

People and continuity. Who will work on the account, what share of their time is ours, and what happens if one of them leaves? The agency selection guide asks this of any agency.

Red flags in a pitch

Any of these can have an innocent explanation. As general practice, press each one with a further question until the reply names something specific.

  • A shortlist of creators the agency represents or is paid by, with no mention of it. Put the representation question from the list above to the agency directly.
  • One price per creator. The negotiation guide asks for a quote broken down by element, and two prices for the same deliverable can cover different things.
  • Follower counts or discovery-tool estimates as the reason a creator is right. The creator selection guide starts from the goal and the audience, and asks for the creator's own insights to be read.
  • Agreements on the agency's paper that leave the brand unnamed. The agents guide wants the brand named as the holder of the content and licence rights, or the route by which they pass to it made clear.
  • Earned media value as the headline result. The ROI guide says to retire it and report the underlying reach and engagement.
  • Links on the agency's own accounts and results from its own dashboard. The agents guide keeps the tracking data in the brand's own analytics, and the ROI guide notes that the comparison with other channels depends on it.
  • Paid amplification promised with no word on the licence or the authorisation. The amplification guide separates the licence from the platform authorisation.
  • Every creator handled as a separate deal. Where the roster is many small creators, the micro-creator tutorial starts from one standard offer.
  • Silence about leaving. A proposal that does not mention a handover has left it to the contract; the exit section of the scope template is where to write it in.

Reading the money

Ask every agency to price the same scope and to split the answer: the agency's own fee, each creator's fee, any licence, product, shipping and any agent commission. The negotiation guide asks for a creator's quote to be broken down by element, since headline figures compare nothing. As general practice the same holds for an agency's proposal: a bundle that hides the split cannot be checked against invoices or go into the CPA that the ROI guide builds from every element, agent commission and agency fees included.

Match the fee model to the work, as the agency selection guide advises, and agree it against the scope. The models it lists, from a retainer to a performance element, each reward something different. Two cautions carry over from the guides as general practice. A model in which the agency's income rises with the creator spend it recommends rewards spending, which is what the agency selection guide says of commission on media spend; fix the ceiling and the split before the model is chosen. And a performance element tied to reach or engagement meets the objection the contracts guide makes to paying creators on those measures, that a creator cannot steer how the platform distributes a post; as general practice the same holds for an agency, so tie any performance element to something the agency can sway and the brand can verify in its own data.

Set the whole against the ceiling. A proposal that fits inside it by leaving out licences, product or the brand's own team time has not priced them; the ROI guide lists team time and licence extensions among the costs that go missing.

Reports the brand can check

A report is the agency's account of its own work, and the brand's records are how it is checked. Match each line below to a record the brand holds.

Report lineBrand's test
Which creators posted, and whenThe roster or tracker rows, the live links and the brand's own look at the accounts
Labels were in placeA sample of live posts read against the label standard in the brief, including every story frame
Visits and orders came from each creator's linksThe brand's analytics, with links built to its own UTM convention, and the shop platform for the orders
Code redemptionsThe shop platform's report for each discount code and the code register, shown on its own line beside link orders, as the KPIs guide's per-creator columns do, because the two methods catch different people; any combined figure should show both parts
Reach and engagement for each postThe creator's insights screenshots in the agreed form, with the dates and the contracted post checked
Cost per outcomeThe brand's own orders and invoices for every element, set against the ROI guide's cost table
Posts backed with paid reachThe rights and authorisation record and the brand's ad account
A return, or an incremental effectWhich part was measured, which estimated and which assumed, as the ROI guide asks of any report

Small numbers need care, and the reading small numbers guide shows why ranking by a rate built on a handful of orders misleads. If the agency's figures climb while the business's stay flat, the flattering agency reports diagnosis sorts measurement gaps from performance gaps.

Judging the agency and the programme

As general practice, judge the agency's work and the programme's results separately. Results depend on the creators, the product and the offer as well as on the agency. The agency's own work can be judged on whether the vetting notes exist and hold up, whether agreements carry what the brand asked for, whether labels are in place when the brand checks, whether reports reconcile and whether the service levels in the scope were met. The influencer programme audit scorecard scores a running programme on rows that match several of these checks. Run it with what the audit's section on access and evidence asks for, naming which lines the agency owns and which the brand keeps.

Judge results against measures written down before the campaign. The KPIs guide asks for one primary KPI, defined before launch, and for data pulled at fixed points; its own example is a pull a week after each post and another at the end of the window, since an early read undercounts posts that keep earning visits. The ROI guide calls link-tracked results a floor and, citing the IPA's effectiveness work, places brand effects over longer spans than most campaign measurement windows. The review date on the one-page strategy is where the judgement is meant to happen.

Do not judge on one post or one creator, and score the brand too, as the scorecard that both sides complete in the agency selection guide does. If the review goes badly, the underperforming agency tutorial begins with the evidence set against the scope, and the pre-renewal agency review fixes the renewal decision ahead of the notice date.

Changing agency, or bringing the work in-house

The sequence for collecting accounts, assets and documentation and for withdrawing access is set out in the agency selection guide's section on transitions and in the agency offboarding checklist, and this page does not duplicate it. Creator work adds points that an account list does not show.

  1. The creator list and the route to each creator. Collect the creator list that the agents guide describes, with the route to each creator, before notice is given. As general practice, settle at appointment what the agency agreement says about the brand approaching creators the agency introduced, and have a legal adviser read that clause.
  2. Agreements and licences in flight. List every agreement with its licence period, its markets and what its termination terms say about unpublished content and fees already paid.
  3. Ads and authorisations. List every ad running through a creator's identity with its licence end date and the status of the platform authorisation. Decide for each whether it continues, pauses or ends, and keep the licence end as a hard stop, as the amplification guide does. Where content has to come out of every place it travelled, the asset withdrawal tutorial traces each use.
  4. Links and codes. The brand's analytics and shop platform should already hold them. As general practice, list every link or code that live posts still send people to and who controls it, so nothing a viewer can still tap or type stops working when the agency's access ends.
  5. Open threads and answers. Hand the answer pack, the threads where a question is waiting and the escalation route to whoever owns them next, as the comments guide describes.
  6. Money. List the orders raised, the invoices approved and the creators awaiting payment, and settle them on the dates agreed.
  7. The creators themselves. Tell them who they deal with next and thank them. The partnerships guide covers ending a creator relationship well.
  8. The pause route. Check that the brand can locate and halt every scheduled post and ad without the agency. The review in the controversy guide asks how long that took and whether access or records slowed it.

If the agency still holds the ad accounts, the ad account handover diagnosis works through why, account by account. If results slip soon after a switch, the post-switch drop diagnosis checks what the handover lost. When the work returns in-house, the agency offboarding checklist notes that the same list applies to the in-house team, and the guide to choosing between an in-house team and an agency covers the general choice.

Weighing agency proposals for a tapestry kit maker (illustrative)

Illustrative example — not a real result

Naidil Tapestry Kits weighs three agency proposals

Naidil Tapestry Kits, an invented seller of tapestry and embroidery kits with a marketing team of 3, plans a launch programme for a new kit line with 9 creators who stitch on video. It writes its one-page strategy, sets a ceiling of £14,000 for fees, product, licences and the agency's own fee, and sends three agencies one brief and one request template, asking for the agency's fee and each creator's fee as separate lines.

Two gates come first: the tracking must run in Naidil's own analytics and shop platform, and every creator agreement must name Naidil as the holder of the content rights or set out how they pass to it. Proposal 1 builds links on its own short-link account and reports from its own dashboard, so it fails the first gate and is dropped however well it might have scored. Proposals 2 and 3 pass. The weights were agreed before anyone read a proposal: vetting evidence 25, contracts and rights 20, disclosure checking 20, reporting Naidil can test 20 and fit with the team 15. Proposal 2 scores 4, 5, 4, 4 and 3 against those criteria, which gives 81 out of 100 once the total is divided by 5, and Proposal 3 scores 2, 4, 3, 3 and 4, which gives 62. Proposal 3 represents 2 of the 9 creators it proposes and says so; Proposal 2 answers the representation question in writing for each name and shows a vetting note for a creator it dropped. Under a sceptical set of weights that puts reporting at 30, disclosure checking at 15 and fit at 10, and keeps vetting at 25 and contracts at 20, Proposal 2 still leads, 82 to 61.

Proposal 2 totals £12,600 for the full brief, inside the ceiling, with the agency's fee, the creator fees, the licences and the product on separate lines. Naidil appoints it for a first round with 5 of the 9 creators, keeps one named reviewer and the final say on must-fix comments, and holds the creator list, the agreements and the licence record in its own files. The agency's first report puts 38 orders on its line headed "orders from creator links". Naidil's analytics counts 31 link orders and its shop platform shows 7 code redemptions, so the report is reissued with each method on its own line and a combined cost per order that shows both parts. Naidil's own read of the licence record finds that two creators' lines carry a licence period but no platform authorisation expiry; the agency asks both creators for the dates and Naidil enters them in its asset library before any ad is launched. Naidil's reads of the vetting notes and of the live labels also pass its own checks, and it keeps the agency for a second round.

Common mistakes

  • Buying the whole programme in one block. A bounded first round tests the three questions set for it, and leaves the brand able to change course.
  • Letting the agency's accounts hold the links, the codes or the ad access. The tracking and the permissions belong to the brand, and an exit without them is a crisis.
  • Comparing bundles. Quotes that are not split by element are hard to compare, check or count in the cost per outcome.
  • Taking the shortlist on trust. Ask who the agency represents, how each creator was checked and where the vetting notes are kept.
  • Treating disclosure as the agency's job alone. The duty can reach the brand too, so the brand reads some live posts itself.
  • Accepting the agency's summary as the result. Read the lines the report is built from, and check that each method has its own line.
  • Judging early, or on one post. Data pulled before the posts have finished earning undercounts them, and one creator is not proof.
  • Briefing nothing. Without the one-page strategy, the campaigns that follow are measured on whichever figures happen to be at hand, as the strategy hub says.
  • Hiring with no exit. The list, the agreements, the rights record and the tracking have to be collectable on the day the arrangement ends.

Frequently asked questions

Should the agency or the brand hold the creator relationships?

The brand should keep the creator list, the agreed terms and a working route to each creator in its own files, even when the agency makes the first contact. The site's guide to creator agents and influencer agencies makes the point that the relationships should stay with the brand if the agency goes. If the agency represents any creator it proposes, ask it to say so in writing for each name, and record the answer beside the shortlist.

Can the agency pay the creators for us?

That is a matter for the scope. Whoever pays, the brand still needs each creator's fee, any product, any licence and any agent commission as a separate line in its own cost record, because the cost per outcome counts all of them. As general practice, raise it while quotes are compared and ask finance to read how invoices will reach the brand's records. Questions about tax forms and paying individuals belong to finance and the brand's advisers.

What if the agency's report and our own analytics disagree?

Treat the brand's analytics and shop records as the reference and ask the agency to show how each line was counted. A report can put code redemptions on the link-orders line, count a post that was not the contracted one, or divide engagement by a different base. Once the lines are separated and the definitions written down, compare the two sources and state whatever gap is left in the report instead of burying it.

Terms used on this page

Frameworks

Sources

  1. FTC Endorsement Guides: What People Are Asking — Federal Trade Commission
  2. Advertising Standards Authority — ASA

Related