TikTok Spark Ads or dark posts for cold prospecting: which one to scale?

Say you're running a DTC skincare brand and trying to scale cold TikTok prospecting. You hit the same fork every time: boost an organic post as a Spark Ad, or push a dark post (ad-only, hidden from your profile) built from the same video. Spark Ads keep the likes, comments and handle visible, which helps trust with a cold audience. Dark posts let you swap copy and CTAs without touching the organic feed, and you don't end up with a comment section full of questions you can't answer fast.

My view is to use Spark for the one or two videos that already hold organic watch time, and dark posts for everything you're still validating. Spark on a flat video just buys you comments under a weak hook. Has anyone seen Spark beat dark posts on conversions, not just CTR?

Agreed on the split. One thing that helps here: Spark Ads carry the organic post's engagement into the ad, so the comments you're worried about are usually already there before you spend anything. If a video has a lot of "where can I get this" or "does it work on oily skin" replies, pay attention to those, because they're exactly the questions your dark-post copy should answer.

Comments only help if someone actually answers them. A cold viewer reads a Spark thread top-down, and an unanswered "does this work on oily skin" looks like a dodge. If you go Spark, reply to the top questions from the brand account before you scale, not after.

Dark posts skip that problem because there's no public thread to manage, which is probably why they often win on tight CPA. The tradeoff is you lose the visible social proof that makes a skincare claim believable. So where does the conversion gap come from, the trust or the copy?

Honestly, I don't think the comment thread is what usually decides CPA. Dark posts save you from managing public questions, but that's an operating benefit, not the main driver. A Spark ad and a dark post built from the same video share the same hook, so if the Spark version converts better, it's often because the organic post already held people past the first three seconds. And if the dark post wins on cost, the copy and offer are the likelier reason, not the missing social proof. Skincare buyers do want reassurance, but a cold viewer still watching at second eight is already past the hardest part. Hiding the comments also throws away a free signal, since those "does it work on oily skin" replies show which objection your dark-post copy should answer first.

I'd push back on the idea that a viewer who makes it past second eight has already done the hard part. For skincare, watching is cheap. The click is where a stranger starts wondering whether this will work on their skin type, and a dark post gives them nothing to check that against. A visible reply like "yes, it works on my oily skin" does that job better than any copy you can write, so Spark can win on CPA even with an identical hook. To tell which effect you're actually seeing, keep the copy and CTA fixed and change only the delivery format.

A visible reply doesn't do the reassurance job better than copy, mostly because you don't control it. On a Spark ad the thread is usually sorted by likes, so the top comments are whatever got the most engagement. For skincare that's often a lukewarm "idk, my skin is weird" or a question from someone whose skin looks nothing like the viewer's. A cold viewer reads that as a mixed signal, not proof. Copy can say "for oily, combination skin, this is the one" every single time.

So I'd expect Spark to win on CPA when the top comments are already positive and lose when they aren't. Keeping copy and CTA fixed, like Rowan suggests, is the right test. I'd just log the sentiment of the top comments next to the CPA, otherwise you can't tell whether the format won or the thread just got lucky.

Agreed, and the thread keeps changing. Early comments on a Spark ad mostly come from people who already watched the organic post, so they're friendlier than what a cold viewer sees once delivery widens. If you only check the top comments on day one, you end up logging a positive thread the cold audience never actually read. Snapshot the top ten comments at launch and again when you read CPA, and note the spend between the two. Otherwise the sentiment column just describes your existing fans.