Which Wellness Drink Brands Rebook Their Creators, and Which Do Not

The category stopped hiring and started rebooking

Here is the shape of the last five years, counted across 71 wellness and functional drink brands with anything on record.

The deal count is still climbing, but the creator count is not climbing with it. A creator who took two drink deals a year in 2021 is taking more than four now, and 2026 is only eight months old.

That is what a category looks like once it has finished experimenting. The testing is done, the winners are known, and the money has moved to a media buy against a list of names that already work.

Who says so: our own database of recorded paid YouTube sponsorships, matched by website domain to a 143-company beverage census. The 2026 row is a partial year, so read the deal count as incomplete and the ratio as directional.

The repeat rate table

Repeat rate here is the share of a brand's creators who came back for a second paid video. Only brands with 15 or more creators on record are included, because a repeat rate calculated on three people means nothing.

Who says so: our own database, counting distinct creators and their deal counts per brand across 30 October 2020 to 7 August 2026.

Two strategies, and only one of them needs an agency

Read the table by its two ends and the category splits cleanly in half.

Ghost Energy, Gamer Supps, EHPlabs and G Fuel run narrow and deep. A few dozen to a few hundred creators, booked eight to twelve times each, until the code is part of the show. Those programmes are mostly a relationship job, and a small internal team can run one well.

AG1, Huel, Magic Mind and Liquid I.V. run wide and shallow. AG1 has 920 creators at 2.63 deals each, which is the largest roster in the category and the lowest repeat rate on the table. Nothing about that is wrong, it buys reach across many audiences at once, and it is exactly how a brand gets to be in every podcast a new customer might listen to.

But 920 relationships is an operations problem before it is a creative one. Somebody has to remember who was paid what, who has usage rights that expire in March, who ran a competitor last quarter, and who quietly stopped answering. This is the specific place where brands call us, and it is the least glamorous work we do. We keep the list clean, we keep the negotiations consistent so the same creator is not quoted two different numbers by two different people, and we make sure a placement is not renewed just because nobody checked what the first one did.

The deepest relationships in the whole category

These are the largest single creator and brand pairings on our record.

Look at the subscriber column. Almost none of these are big channels. The deepest relationship in the category belongs to a fitness creator with 123,000 subscribers who has read the same electrolyte code 117 times.

That is the clearest argument in our data against buying on follower count. LMNT did not win by renting the biggest audiences, it won by finding people whose audience actually buys and then never letting go of them. We wrote up how that programme is built in our LMNT campaign analysis.

Who says so: our own database, counting deals per creator per brand. Subscriber counts are the most recent figure we hold for each channel and move over time.

A repeat rate is the closest thing to proof that exists

Almost every number in influencer marketing measures attention. Views, impressions, engagement rate, earned media value, all of them describe how many people saw something.

A repeat rate measures a decision. Somebody at that brand looked at what the first video did and chose to pay for a second one. Nobody rebooks a placement that lost money 85 times in a row.

That makes this table more useful than a case study, because it is not the brand talking about itself. It is what the brand actually did with its own money, read back out of the public record.

It is also the cheapest research you can do before an outreach campaign. If you are about to email a creator, knowing they have taken 40 deals from a competitor tells you your pitch is not landing on neutral ground, and knowing they have taken one tells you the opposite. The gaming brands that sit at the top of this table are the subject of our post on who really buys energy drink creator deals.

What a low number does not mean

Three honest limits on this table.

We read sponsorships out of YouTube video descriptions, so brands that ship with a promo code look enormous here and brands that sell cans through a shop look quiet. Red Bull at 19 recorded deals is not a small spender, it just spends on athletes and events, which leaves nothing for us to count.

We hold no TikTok or Instagram data, so a brand running its whole programme there reads as a low number or a zero.

And a first-year brand will always show a low repeat rate, because it has not had time to rebook anybody. Bloom Nutrition at 33 percent on 18 creators is a young programme, not a failing one.

What we would do with this

If you are building a roster, copy the deep model first and the wide model second. Find twenty people whose audience matches, book them four times each instead of eighty people once, and let the code do the work of proving which ones to keep.

If you already have a wide roster and it has stopped being manageable, the fix is not more creators. It is a single list that says what each person cost, what their code did, and when their rights run out.

Before any of that, screen the people you are about to approach. In a health and wellness category the reputation check is not optional, and the claims a creator makes on your behalf are your legal problem, not theirs. Our functional beverage vetting playbook covers the four archetypes and the questions we ask on the call.

We find and vet these creators, we check what they have already run and what they charged, we negotiate, and we keep the roster honest so a renewal is a decision rather than a habit. If you sell a drink and want to know who on this table is reachable this quarter, that is the conversation to have.