Avatar Algorithm

An avatar algorithm is a discovery method that treats the For You / Explore feed as the source of truth. You define an ideal customer (the avatar): niche, age band, the problem they are trying to solve. Then you scroll TikTok and Instagram as that person, on real phones, and record the creators the platform keeps serving. The output is not “everyone who used #habitapp.” It is “who is trending in the feed your customer actually sees.”

That distinction matters because social graphs and hashtag indexes describe the past. Feeds describe what the ranking systems will put in front of a similar person tomorrow morning.

Why it matters to an indie iOS or SaaS founder

If you shipped a live app and downloads are flat, you already know Apple Search Ads and a Product Hunt launch are not a growth engine. The remaining path that still looks native to a consumer is a creator talking into a camera about the same pain your screenshot five tries to explain.

The failure mode is picking the wrong person. A 400k-follower “productivity bro” will quote a rate your MRR cannot survive, then post a video whose audience already owns Notion, Todoist, and three other trackers. You paid for reach into a feed that is not yours. Hours of manual scrolling produce the same error at a slower pace: you open profiles, guess from bios, save a spreadsheet, and still have not seen the account the way a 26-year-old with screen-time guilt sees it.

An avatar algorithm is a way to stop guessing from the creator’s side of the market (who wants brand deals) and start reading from the customer’s side (who keeps showing up in their session). For a solo founder, that is the difference between one expensive miss and a shortlist of people the ICP already trusts.

How feed-as-avatar discovery actually works

You cannot “scrape TikTok” in the abstract and get a useful list. The interesting object is a session: a warmed device, a niche, a language, a pattern of watch-time. Two phones that both follow “fitness” will diverge after a few dozen videos because ranking is personalized. That is the point. You want the personalization, not a global crawl.

A practical avatar is narrow. “Women 18–34” is a media plan, not an avatar. “Someone who already tried three habit apps, hates streaks, and watches ADHD and morning-routine videos” is an avatar. You tell the scroller what to lean into. The feed responds. Creators who recur, same face, rising views, comments from people who sound like users, go on the list.

What you capture is boring on purpose: handle, rough niche, whether they already talk about apps or adjacent problems, a public contact if one exists. You do not need a full media kit. You need enough to decide “this person is in the feed” versus “this person bought a hashtag.”

How JumpTrend approaches it

JumpTrend’s product is this method. You describe the avatar. JumpTrend runs real phones and scrolls TikTok and Instagram in that niche. You get the creators trending in that feed. Hashtag tools and influencer databases are the thing it is built against: they tell you who exists and who used a tag. JumpTrend tells you who the algorithm is pushing to your customer.

Every plan includes avatar setup and the scroll. Scout ($1,500/month) stops at the list, you take it from there. Outreach ($2,500) contacts those creators and follows up until they reply. Dealmaker ($5,000) adds contract generation. Creator fees are never in the plan price; you approve every creator and every rate. There is no demo and no sales call. Scrolling starts once the avatar is defined.

JumpTrend does not guarantee views or downloads. It guarantees the work it controls: relevant creators pulled from the avatar’s actual feed. Narrow niches are not a bug. They often have loyal audiences and rates that have not caught up to reach.

If you already have a growth team and a Modash seat, this is probably not your tool. If you can ship but cannot spend evenings in DMs, the avatar algorithm is the part of marketing that should not live in your head.