Every Group Filter in Cruva, Explained

When you build a group in Cruva, you're really building a set of rules. Every time the group refreshes, Cruva checks every creator connected to your shop against those rules, and whoever matches is in the group. This article walks through every filter available, section by section, and explains exactly what each one does.

Every Group Filter in Cruva, Explained

When you build a group in Cruva, you're really building a set of rules. Every time the group refreshes, Cruva checks every creator connected to your shop against those rules, and whoever matches is in the group. This article walks through every filter available, section by section, and explains exactly what each one does.

What Are Groups?

Groups are dynamic segments of your CRM. Rather than being static lists, they continuously track every affiliate that matches the conditions you define: as creators start or stop matching, the group updates on its own.

That makes groups the engine behind automated outreach. Attach a group to a CRM campaign and it will message creators the moment they hit a milestone or interact with your brand in a qualifying way, with no manual list uploads required. As new creators flow into the group, the automation springs back into action and picks them up.

Groups aren't only for outreach, though. You can use them to track data on a segment over time, or flip them around and use them as an exclusion, keeping everyone in a given group out of a campaign entirely.

How Filters Work

One important concept before we start: some filters look at a creator's relationship with your shop (how much they've sold for you, when they last posted about you, whether you've messaged them), while others look at the creator's overall profile across TikTok Shop (their followers, their engagement, their total sales across every brand they work with). We'll flag which is which as we go.

Samples & Products

These filters are about the free samples you've sent creators and what happened after.

Sample Request Status — Filters creators by where their most recent sample request sits in the pipeline: requested, shipped, delivered, content pending, posted, overdue, and so on. There are two special statuses worth knowing. Ignored pulls in creators whose sample requests you've marked as ignored. No Request flips the filter inside out: instead of finding creators with a matching sample request, it finds creators who have never made one. That makes it perfect for cold-outreach groups — "everyone we haven't sampled yet."

Sampled Products — Only include creators who requested a sample of specific products you choose. When combined with Sample Request Status, only samples of these products in those statuses count. A dropdown to the left of the product selector controls the matching logic:

  • OR (default): the creator has sampled at least one of the products in the list.
  • AND: the creator has sampled every product in the list (and possibly others too).
  • Only: the creator has sampled at least one product in the list, and nothing outside it.
  • NOT: the creator has not sampled any of the products in the list.

Exclude Sampled Products — The mirror image: kick out any creator who has ever requested a sample of the products you list. Useful for "haven't tried our new flavor yet" style targeting.

Sampled SKU IDs — Same idea as Sampled Products but at the variant level. If you sell a serum in three sizes and only want creators who sampled the travel size, this is the filter.

Sample Request Source — Filters by how the sample request came about — whether it originated through a Target Collab or an open collaboration. Only those two sources are accepted.

Sample Received — A day-range filter counting from the moment the creator's sample was delivered. "At least 7 days, at most 30 days" finds creators who got their package one to four weeks ago. Great for timed follow-ups: enough time to have tried the product, not so long that they've forgotten you.

Note: All of the sample conditions combine into a single logical clause applied to the same sample. For example, Sampled Products = Y plus Sample Request Source = open collaboration means the creator must have sampled product Y through an open collab — not product Y in one request and an open collab in another.

Content Unfulfilled — This one is a workhorse for chasing content. It finds creators who received their sample a certain number of days ago and still haven't posted — technically, creators whose sample is sitting in "Content Pending" or "Overdue" status. "Content unfulfilled for at least 10 days" is the classic gentle-nudge automation trigger.

Showcase Status — Whether the creator currently has your product in their TikTok showcase. A creator showcasing you has your product one tap away from their audience, so this is a strong intent signal.

Previously Worked With — A simple yes/no. "Yes" means the creator has at least one sample that made it to a meaningful stage — shipped, posted, or in the content pipeline. "No" means they've never gotten that far with you. This is the cleanest way to split your CRM into "existing partners" versus "brand new prospects."

Commission Range — Filters by the total commission the creator has earned from your shop. High earners are your proven partners; zero earners might need a different pitch.

Content

These filters are about the posts creators have made for your shop.

Posted Products — Only include creators who have actually published content featuring specific products of yours. Note the difference from Sampled Products: sampling means they asked for the item; posting means they made the video. A creator can appear in one and not the other.

Number of Posts — How many videos the creator has posted for your shop, as a minimum/maximum range. Use it to find your most prolific posters, or the one-and-done creators worth reactivating.

Days Since Last Post — How long it's been since the creator last posted about your shop. A minimum of 30 days finds creators who've gone quiet; a maximum of 7 days finds who's actively posting right now. This filter is the backbone of most re-engagement groups.

Performance

These are the money filters.

Shop GMV — Total sales the creator has generated for your shop specifically, as a range. This is your own data about your own relationship with them.

30d Affiliate GMV — The creator's total sales over the last 30 days across every shop they promote, not just yours. This is the filter to use when hunting for talent: a creator might have sold nothing for you yet but be moving serious volume for other brands. Pairing "Shop GMV is zero" with "30d Affiliate GMV is high" finds exactly the creators worth pursuing hardest.

Units Sold — Total number of individual units the creator has sold for your shop. Useful when your products are low-priced and GMV understates how much a creator actually moves.

Lives

Live streams behave differently from videos, so they get their own section.

My Shop LIVE GMV — Sales the creator has generated for your shop specifically through live sessions, as a range. Some creators barely post videos but move real volume live; this filter finds them.

LIVE Count — How many live sessions the creator has run for your shop. (There's also a live duration range available — total time spent live — for finding the true marathon streamers.)

Days Since Last LIVE — How long since the creator last went live for your shop. Just like Days Since Last Post, but for streams: a big minimum finds lapsed live sellers, a small maximum finds who's live right now.

Creator Profile

These filters mix the creator's public stats with your communication history.

Followers — The creator's follower count, as a range. Standard tiering: nano, micro, mid, macro — slice it however you like.

Engagement Rate — The creator's average engagement on their content, as a percentage range. A creator with 5,000 followers and sky-high engagement often outsells one with 500,000 followers and a dead audience.

Post Rate — How reliably the creator actually posts after receiving samples, as a percentage. This is one of the most underrated filters in the product: a high post rate means that when you send this person a sample, content actually shows up. Filtering for a strong post rate before sending samples saves a lot of wasted product.

Has Email — Whether Cruva has an email address on file for the creator, from any source. Essential when building groups destined for email outreach — no point targeting people you can't reach. It can also be flipped to find creators missing an email, if you want a list of contacts to enrich.

Has Phone — Same idea, for phone numbers. "Yes" requires an actual, non-empty number on file.

Has Replied — Whether the creator has ever replied to a message from your shop. "Yes" gives you your warm list — people who've engaged in conversation. "No" gives you everyone who's gone silent or was never contacted.

Days Since Last Messaged — How long since anyone at your shop last messaged this creator through Cruva, measured from the most recent message only. A minimum of 14 days means "we haven't touched this person in two weeks" — the natural guardrail that stops automations from spamming the same creators over and over. (There's also a simpler "has been messaged at all" version of this check.)

Days Since Last Replied — How long since the creator last wrote back. A small maximum finds creators with fresh, active conversations that deserve a human follow-up; a large minimum finds conversations that have gone cold.

Messaged By Campaign — Filters creators by which of your outreach campaigns have messaged them. This one supports the same style of matching logic as Sampled Products. The basic version includes anyone messaged by at least one of the campaigns you select. There's an option to narrow it to creators who have only ever been messaged by those campaigns and no others, an option requiring the creator to have been messaged by every campaign on your list, and an inverse — creators who have not been messaged by certain campaigns — which is how you keep separate outreach tracks from stepping on each other.

Demographics

All of these describe the creator themselves, based on Cruva's analysis of their content and profile. Each one is a multi-select: pick several options, and a creator matches if they fit any of them.

Gender — The creator's gender.

Age Range — The creator's age bracket (e.g., 18–24, 25–34).

Race/Ethnicity — The creator's race or ethnicity, useful for representation-conscious casting.

Body Type — The creator's body type — commonly used in apparel and swimwear, where fit and representation drive conversion.

Economic Status — The creator's perceived economic presentation — whether their content reads as budget-conscious, mid-market, or luxury. Matching this to your price point matters more than most people expect.

Content Tone — The overall style and vibe of the creator's content: funny, educational, aesthetic, high-energy, and so on.

Language — The primary language of the creator's content.

Alongside these, you can also filter by the creator's content categories (beauty, food, fitness, etc. — a creator matches if any of their categories overlap with your picks) and by their audience: the age brackets that make up the creator's top follower segments, so you can target creators whose viewers are 25–34 rather than creators who are.

Demographic filters also support advanced boolean grouping: you can build multiple condition sets where everything inside a set must be true together, and a creator qualifies by matching any complete set. For example: ("female AND 18–24 AND beauty") OR ("male AND fitness"). Within a single simple filter panel, all your chosen filters must hold at once.

Tags

Include Tags — Only include creators carrying at least one of the tags you select. Tags are your own labels — "VIP," "holiday-campaign," "do-not-rush" — so this filter is how your team's manual organization becomes targetable.

Exclude Tags — Kick out any creator carrying any of the selected tags. The classic use is a "do not contact" or "already in another campaign" tag that keeps creators out of automations regardless of what else they match.

Was this article helpful?

Related articles