IMOKTide
Sign in

Picking creators takes a method.

Method

Picking creators isn’t about who has the most followers. Tide breaks the call into four simple moves — you can follow every one and change it.

01Peer rank

Only compare creators of a similar size

Big accounts naturally get lower engagement rates and small ones higher, so putting them side by side isn’t fair. We only compare creators of a similar size (within 10×, same source, time window and content type) and skip the rank if there are fewer than 10.

02Health

Unhealthy accounts get flagged

The platform grades account health every month. Abnormal accounts often have the best-looking numbers — we flag them in red so you can drop them in one tap.

03Selection plan

One waterline is one condition

“Engagement rate better than 75% of peers” is a waterline. Drag it, and who floats to the surface changes right away.

04Budget mix

Give us a budget, get the best mix

Tell us your budget and how many creators at most. Tide works out which group in your shortlist gets you the most reads (or engagement, or people) for that money. It’s a suggestion — you decide.

Screening score · demo data

How the screening score works

Every row on the board has a screening score from 0 to 100. It isn’t a black box: how much each of the four parts counts, how far each one beats the group, and which number it used — it’s all laid out.

  • Scale 30%Follower count
  • Activity 20%Posts in the last 30 days; if unknown, days since the last post
  • Performance 30%Image-and-text platforms: median engagement (or engagement rate). Video platforms: median views
  • Region 20%The creator’s stated location vs your target region: match 100, mismatch 0, can’t tell = not assessed
  1. Compared only within the same platform and target region: every part is a 0–100 percentile in that group.
  2. Missing never counts as 0: that part is “not assessed”, the total is reweighted over the rest and marked “provisional”.
  3. Fewer than 2 parts assessed means no score: “not enough evidence”, priority “to verify”.
  4. Groups with fewer than 5 creators get “small sample”: the score still counts, but can’t reach P1.
  5. P1 = total ≥ 70, region matches, no hard risk, not a brand account; P2 = total ≥ 50, no hard risk; a region mismatch goes straight to P4 on hold.
  6. Risks are noted, never deducted: unavailable posts, no recent posts and abnormal health only hold back the priority — the score doesn’t move.
See the full demo on the Board

Score breakdown · Sample IG-JP-10

89.4ProvisionalP2

  1. Scaleweight 30% → actual 37.5%Followers · 721Kbeats 100% of the group14 comparableOKadds 37.5
  2. Activityweight 20% → actual 25%Posts, last 30 days · 20 postsbeats 69.2% of the group13 comparableOKadds 17.3
  3. Performanceweight 30% → actual 37.5%Median engagement · 16.6Kbeats 92.3% of the group13 comparableOKadds 34.6
  4. Regionweight 20% → actual Not assessedStated location vs target region · No stated locationNot assessedNot assessed—

Σ contributions = 89.4; 3 / 4 parts assessed, reweighted over 80%

Group = Instagram × Japan, 14 creators, ranked #1

  • To verify · Region to verify

Risks are noted, never deducted. · rules screen-v0.1

Examples come from the Board page’s demo data, scored live by the same function.

Peer bar

Two kinds of “peers”, each with its own job.

Both kinds only compare inside their own group — never across groups.

01Selection pool · rank

Creators of a similar size

When you pick, you look at rank: creators are only compared with others within 10× their follower count, and there’s no rank if there are fewer than 10. Big and small never share a bar, so “better than 75% of peers” is fair to everyone.

02Board · screening score

Same platform, same target region

When you screen, you look at groups: creators on the same platform for the same target region form one group, and percentiles are worked out inside it. Instagram creators for Japan never get compared with Xiaohongshu creators for Taiwan. Groups under 5 are marked “small sample”.

Health check

Abnormal health: flagged red, held back, not deducted.

The platform grades account health every month. Abnormal accounts often have the best-looking numbers, so we flag them red on the spot with the reason, and you can drop them from the selection pool in one tap. On the board it counts as a hard risk: it can’t be P1 or P2.

Why not take points off?

Deducting would mix “unhealthy” and “just average” into one number, and you couldn’t tell which it was. The score only speaks to performance; risks sit right next to it where you can see them.

Three hard risks

  • Abnormal health
  • Posts unavailable (private or failed to load)
  • Nothing posted in the last 30 days

Demo data · up to 10 creators

Give it a budget, get the mix that goes furthest.

Name a budget and we'll pick the mix that buys the most reads with it. It looks at the one goal you choose (reads here) — it never blends several into one score. Quotes use the lowest price, and it's only a suggestion: you decide whether to take it.

10 creators · ¥100K · about 229K reads

Demo quotes are made up from followers and engagement, rounded to ¥100; accounts flagged unhealthy sit out.

Demo: 117 healthy sample creators (11 flagged unhealthy sit out). With a ¥100K budget and at most 10 creators, the suggestion is 10 creators for ¥100K, about 229K reads.

Tap the water

Start a shortlist you can stand behind.

Just sign in with your work email — we’ll take you straight to the workspace for your role.

Sign inRequest access