Creator Discovery for Agencies

I find creators for an agency client by writing one avatar per brand, scrolling TikTok and Instagram as that customer, overlap-checking anyone already on the roster or last quarter’s deck, and returning 8-25 handles with one video each. Not a follower-banded spreadsheet. Not last campaign’s leftover yeses.

“Women 18-34, wellness, US” is how two shops produce the same forty names. If the client cannot describe the customer in a sentence a phone could learn, charge for the brief. Do not paper it with a database export.

This is not a replacement for Modash, Aspire, or Grin. Those are indexes and CRMs. You still need history and payments. You needed a different filter for the first slide: in this client’s session, not in the category. JumpTrend’s agency surface is jumptrend.co/agencies.

One avatar per client

You already have clients, or you are about to. You will not scroll a fresh For You for every brand every week and still bill hours. So you fill the deck with whoever replied last time. That is rational. It is also how retainers get compared on price.

Feed-as-avatar discovery is a way to walk in with names the client’s customer is being served this week, not names that matched a niche checkbox last quarter.

Each plan, each scroll, is one avatar. Do not run five brands on one phone. The feed will smear. Rotate a test avatar on Scout if you must; do not pretend a blended session is five deliverables.

Artifact: the client-avatar brief

This is the shape you can put in a kickoff doc. It is not a capabilities slide.

Client: [brand]   Product: [app / SKU / both. say which]
Destination: App Store / Play / Shopify / mixed. If mixed, this avatar is for: [one]
Customer sentence: [problem, quit-list, adjacent videos, language, geo]
Do not include: [the client's other persona they will ask for "while you're in there"]
In-feed definition: creator recurs in a warmed TikTok/IG session for that sentence
Out of scope: emails, rates, Spark, the last campaign's leftover yeses
Overlap check: flag anyone already on our roster or in last quarter's deck
Deliverable: 8-25 handles, one video each, public contact if any
Whose name on outreach: ours / the shop's  (decide before anyone emails)

The overlap line is the agency-specific part. See creator roster overlap. A feed-qualified name that three of your clients already burned is not a fresh deck. It is a CRM query wearing a new PDF.

What you keep

Scout is $1,500/month for one list. You pitch, you negotiate, you keep the retainer. Outreach is $2,500 if we go out in your name or JumpTrend’s; you stay the face of the account. Dealmaker is $5,000 for paperwork you still approve. Creator fees sit with the client, not inside the JumpTrend invoice. No demo. No download guarantee.

Most shops should start on Scout and keep inbox in-house. Use Outreach or Dealmaker when the client wants motion faster than your roster allows. One extra avatar is the usual first buy, not a six-seat creator OS.

JumpTrend will not talk to your clients on Scout. That is the point.

What not to sell as the deliverable

A follower-banded spreadsheet (“15k-50k fitness”)
A reply-rate
A CPI
“We found everyone in the niche”
The same ten mid-tiers with new screenshot crops

Sell: a session-specific shortlist, overlap flagged, one video per name so the client can do the stupid test (“would my user know them?”).

How this sits next to the database you already pay for

Keep the database for: emails, audience estimates, fraud hints, campaign history. Use the avatar scroll for: who is actually on this customer’s For You. Paying two seats to export the same CSV twice is how discovery becomes a tool tax. Paying a $1,500 avatar so the first slide is not last quarter’s leftovers is a different invoice.

If the client is Shopify-native and already lives in Aspire or Grin, do not rip that out. Add a feed-qualified set for the app-install workstream, or for the one persona the marketplace keeps missing.

Measurement, because the client will ask

You still will not get Shopify-grade affiliate truth on an iOS IAP. Do not put platform view-through in a bonus clause. See Measuring creator UA without vanity views. Discovery quality makes a later lift more interpretable. It does not make a lift automatic.

The free brief tool at jumptrend.co/brief is usable as a client intake if you do not want them in your Notion.

Neighboring: Creator deck, Boutique creator agency, Avatar algorithm.

When the client is an app and you are a Shop-shaped shop

Many boutiques grew up on unboxings. An app client will still ask for a deck. Do not import a gifting workflow. There is no parcel. There is a TestFlight and a one-pager. If your existing CRM assumes SKUs, keep it for the DTC clients and run a thinner process for the app: avatar, list, brief, rate, usage. JumpTrend will not become your Shopify fulfillment.

If the client is mixed (app plus merch), one avatar per destination. A blended feed produces names you cannot brief.

What “first extra avatar” means

You already have a house process. The first JumpTrend plan is usually one additional client session: the brand whose last deck looked like everyone else’s. Not a rip-and-replace of Modash. If that session is useful, add a plan. If it is not, cancel. Monthly billing exists so this is not a religious conversion. One avatar per client. Then you stop recycling the same ten names because they reply.