Manifesto

The influencer industry runs on someone remembering

Micro and nano creators are the ones who actually sell. The operations around them are held together by attention, and attention does not scale. Here is the whole sequence we are building, including the parts we have not built yet.

Small creators sell. Big ones stopped.

Micro and nano creators are the ones who actually move product.

The mega influencer era arrived fast and left just as fast. Paying one person six figures for a single post stopped working, and most brands already know it. What replaced it is not smaller in total. It is spread across far more people.

YouTube is producing new creators every month. TikTok and Instagram keep surfacing people with two thousand followers who sell out a product because their audience actually believes them.

A recommendation from someone who looks like your neighbour beats a recommendation from someone who looks like a billboard. People buy from people they think genuinely use the thing.

Here is the catch. Working with 500 small creators is 500 times the operational work of working with one big one. Same revenue, far more admin. That is the real reason brands still overpay one person instead. Not because it works better. Because it is easier to manage.

The whole thing is one person holding it in their head

One influencer manager is tracking dozens of creators across a DM inbox, a WhatsApp thread, a spreadsheet, and their own memory. Sourcing happens in one place, addresses in another, follow up in a third, and the record of what actually got posted usually nowhere.

Nobody involved is lazy. The process is just held together by attention, and attention runs out. So the same things go wrong every time:

  • Creators get missed because a message scrolled out of view.
  • Nobody knows the real ratio of products shipped to videos posted.
  • A creator deletes the post after three days and no one notices.
  • Rights to use the video in ads were never signed, or were signed in a thread nobody can find.
  • The end of campaign report is assembled by hand, from memory, days late.

Every one of those is a manual step that failed. More manual steps means more mistakes. That is a systems problem, not a talent problem.

An operations engine, not another database

We are not building a marketplace. We are not building a discovery tool you pay two thousand dollars a month to search.

We are building the unglamorous machine that runs the whole sequence, from finding a creator to confirming their video is still live and paying them for it.

It is simple to describe and painstaking to build, because every single step is a place where a person currently improvises. We are replacing the improvising with a workflow, one step at a time.

Here is the entire sequence. Some of it is live today. Most of it is not. We will say which is which.

01

Discovery

Building

Today. Someone scrolls, copies handles into a sheet, and hopes they are not contacting people they already burned last month.

What we are building. One place every creator gets pasted or pulled into, scoped per brand, deduped against everyone you have ever contacted. Unexciting, and it gives back an hour a day.

02

Vetting for quality and aesthetics

Building

Today. You open every profile and make a judgement call. It is the slowest part of the job and the one you cannot hand to anyone, because taste is the actual thing you are protecting.

What we are building. An engine that learns what fits a specific brand and hands back a short list, so the judgement call becomes a review instead of a search.

This is the hardest thing on this page and we will be straight about it: it is not solved. Follower count and engagement rate are not taste. Everyone selling you filters is pretending they are.

03

Reaching out

Running

Today. Manual emails and DMs, sent when someone remembers, from whichever inbox happens to be open.

What we are building. Sequences that send from the brand's own mailbox, warmed properly so they land, with every reply sorted as it arrives. You look at counts, not at an inbox.

04

Following up

Running

Today. The part that quietly gets dropped. Most creators who never reply are not saying no. They just did not see it.

What we are building. Follow up that runs on a schedule and stops the second someone answers. Nobody falls out of the pipeline because a person got busy.

05

Sending the product

Live

Today. Someone copies an address into Shopify by hand, picks the variant, and hopes nothing got typed wrong.

What we built. One branded link. The creator picks their product and their variant and types their own address. A real zero dollar draft order lands in your Shopify admin, ready to fulfil. No coupon codes, no spreadsheet, no copy and paste.

This is the piece that already works, and it is the reason the rest of this page is worth reading. We shipped one step properly before claiming the other seven.

06

Getting the video back

Building

Today. You ship the product and then you wait. A large share of what you ship never becomes anything, and most brands cannot tell you what share that is.

What we are building. Follow up that runs itself across email, WhatsApp, and calls, plus the brief attached to the gift so the creator knows exactly what to make.

Most creators who go quiet are not dodging you. They are stuck on what to film. Hand them a brief and the content appears.

This one number decides whether creator gifting makes money: how many shipped products turn into posted videos. We measure it. Almost nobody does.

07

Verifying the video

Not built

Today. Nobody checks. On a campaign with hundreds of videos there is no person whose job it is to confirm each one went up, stayed up, and did what was agreed.

What we are building. Automatic checks. Did it post. Is it still live a week later or was it quietly deleted. Was the product on screen for as long as agreed. Did the audio cover what the brief asked for. Does it match the look that was approved.

Not built yet, and we are not going to pretend otherwise. Verification bolted onto a leaky pipeline verifies nothing.

08

Paying

Not built

Today. Paid by hand, tracked in a sheet, argued about a month later.

What we are building. Payment tied to what actually happened. Flat per creator if you want it simple, or per thousand views if you would rather pay for outcomes. Money releases after the post has stayed live for the agreed window, not before.

What is live, and what is not

Step five is live and running for real brands on the Shopify App Store today. Steps three and four run every day for the brands we operate directly. Steps one, two, and six are being built now. Steps seven and eight are not built.

We are publishing this list before it is finished on purpose. Every creator tool we have used oversells the parts that do not exist, and we would rather you know exactly which square of the board we are on.

We run this on our own store first

Sumeet runs growth at Dr Water, a Shopify brand doing about two million dollars a year. Seed was built to run Dr Water's own creator gifting before it ran anyone else's, and every feature still ships there first. If it does not survive our own campaigns it does not go out.

We are two people. There is no sales team and no account manager. Seeding costs two dollars fifty per gift shipped. Verification will be priced separately once it exists, because checking a video is worth more than sending a box.

We are not trying to have the most features. We are trying to build one sequence that runs all the way through without a person holding it together.

The boring version of this is the whole product

If you are shipping product to creators and losing track of it somewhere between the parcel and the post, that is the problem we are working on.