Boutique Influencer Agency Workflow for Mobile Apps

I run an app client at a 3-15 person shop like this: refuse a fuzzy kickoff, overlap-check the roster, scroll one avatar on one session, watch a video per name, deck 8-25 handles, get a client yes on names and a rate band, then brief, then email, then ship TestFlight the same day. There is no parcel. There is a listing, a one-pager, and a destination that is not a product tag. Do not import a gifting calendar.

The deliverable is not “we found influencers.” It is a session-specific shortlist the client can reject, a brief a creator can film from, rates the client approved, and a post that points at a store, not a Shop SKU. You already have a roster. The app-shaped failure is using it anyway because it is in the CRM.

JumpTrend is matching (and optional inbox/paper). You keep the retainer. See Boutique creator agency.

Refuse a fuzzy kickoff

Customer sentence, quit-list, language/geo, destination, whose name is on outreach. If they cannot say the sentence, charge for the brief. Do not paper it with a database export.

Write the destination in the kickoff: App Store / Play / mixed. If mixed, this workflow is for one avatar. A blended feed produces names you cannot brief. UGC for mobile apps is the artifact shape; app influencer seeding is how access replaces a mailer.

Overlap pass before anyone scrolls

Flag handles already on this client’s last deck, on two other clients, or burned. A feed-qualified name that three retainers already used is a CRM query wearing a PDF.

One avatar, one session

Scout is $299/month, or a disciplined DIY scroll. Not five brands on one phone. JumpTrend method: real phones, TikTok and Instagram, as the client’s customer. Not unauthorized access; not a hashtag dump.

Safety and craft pass on the shortlist

Watch a recent video per name. In-feed is not brand-safe. Follower count is not a sort key.

Deck: 8-25 names

One video each, why they are in this feed, overlap flagged, public contact if any. No reply-rate column. No CPI.

Client yes on names and on a rate band

Before anyone is promised money. Talent is extra on every JumpTrend plan. You do not hide a markup inside Scout.

Brief after the yes

One page: mechanic to show, claims they must not make, install path (code / link / TestFlight), due date, disclosure, usage (organic vs Spark).

Email, then provision access

Outreach in the shop’s name unless the SOW says otherwise. Outreach is $499 if JumpTrend should send and follow up; you stay the face. Contract is $1,500 for paper you still approve. No demo. No download guarantee.

Unique codes or seats. Empty-state sample data if the UI is sad. Answer product questions like an adult; seeded creators hit your worst flow.

Post, then measure the decision you wrote down (repeat / Spark / stop). Views are not UA. Do not put platform view-through in a bonus clause.

Mistakes that waste the week

• Running the Shop workflow (samples, product tags, GMV dashboard) on an IAP.
• Filling the deck from whoever replied last time.
• Selling “we found everyone in the niche.”
• Quoting a fake ROAS to win the kickoff.
• Letting the client add a second persona “while you’re in there.”
• Paying two seats to export the same CSV (database + avatar scroll used as duplicates).
• Treating JumpTrend as a UA agency. It does not buy media. See JumpTrend vs hiring a UA agency.

If this calendar slides because Slack won, you did not have a workflow. You had a mood.

How JumpTrend’s method changes the list (not a download promise)

Hashtag and lookalike tools answer who exists. A warmed session answers who this customer’s For You is serving this week. That is the only thing Scout is for.

Keep the database you already pay for: emails, audience estimates, fraud hints, history. Use the avatar for the first slide. Most shops should start on Scout and keep inbox in-house. Use Outreach or Contract when the client wants motion faster than the roster allows. Cancel if the session is not useful; monthly billing exists so this is not a religious conversion.

JumpTrend will not talk to your clients on Scout. That is the point. It will not guarantee views or downloads. Matching quality makes a later lift more interpretable. It does not make a lift automatic. One avatar per client, not one blended phone. Then you stop recycling last quarter’s mid-tiers because they were easy.

What to measure without inventing ROAS

Pick the decision before the post: repeat this creator, put Spark behind the hook (only if the contract allows), or stop. Log unique codes redeemed, store traffic you can actually see, and whether a real user would know the handle.

Do not report a CPI you cannot defend. Do not bonus on vanity views. If the client is mixed (app plus merch), do not blend the scoreboard.

Intake you can steal: jumptrend.co/brief. Product entity: JumpTrend.