The path to purchase almost never looks like a straight line anymore. It's closer to a tornado: a customer bounces through half a dozen different ads, in no particular order, before ever pulling out a credit card. Ads Manager, though, still reports on that journey as if it were linear. Whichever ad happens to get clicked right before the checkout event gets credited with the entire sale, and every ad that came before it, the ones that actually built the desire to buy, gets nothing.
That creates a real blind spot for growth teams. If you're judging every top-of-funnel (ToF) video ad purely on its direct cost per acquisition, you're likely turning off the exact campaigns responsible for filling your pipeline in the first place. To see the true impact of top-of-funnel creative, you need two metrics most accounts never look at: Ad Saves and Ad Shares, and the Cost per Save and Cost per Share that come with them.
The Last-Touch Attribution Trap
Picture a prospective customer scrolling in bed who comes across a genuinely engaging, educational video ad. They're not ready to buy in that moment, they're commuting, working, half paying attention, but the creative resonates enough that it sticks. Days later, after being retargeted several times, they finally convert, on a basic bottom-of-funnel (BoF) static image offering 10% off.
Ads Manager assigns full credit for that purchase to the static discount ad. The video that actually did the work of educating the customer and building their intent to buy shows up as a high-CPA underperformer. A media buyer, reasonably, turns it off to protect efficiency. A few weeks later, the account stalls, not because the discount ad stopped working, but because the thing feeding it new, warmed-up prospects got shut off.
Ads Manager only sees the last click. The video ad that built demand over multiple touchpoints receives no credit at all, even though it did the actual selling.
Why Saves and Shares Indicate Real Intent
Saves and shares are the closest thing Meta gives you to a customer raising their hand. A view can happen by accident, a thumb pauses mid-scroll for half a second and the platform counts it. A save or a share requires an active decision. When someone saves an ad, they're telling you, in a completely unprompted way, that they intend to come back to it. When they share it, they're pulling a second decision-maker, a partner, a friend, a group chat, into their own buying process.
Neither of those actions is something a person does for an ad they don't care about. That's exactly why they're such a clean signal for ToF creative specifically: they measure whether the ad actually landed, independent of whether the viewer happened to be in a position to buy in that exact moment. By mapping out Cost per Save and Cost per Share, you can accurately benchmark which ToF creatives are genuinely hooking your target audience, even when the direct ROAS looks unimpressive on day one.
This lines up with where Meta's own delivery system has been heading, too. We've written before about how Meta's Andromeda system has made creative diversity the new targeting lever, and engagement depth, not just clicks, is a meaningful part of how the algorithm decides what to keep showing. An ad that earns saves and shares is an ad the algorithm has more reason to keep serving, on top of the demand-building effect it's having on the humans actually watching it.
How to Actually Add This to Your Reporting
This isn't a hidden or hard-to-reach metric, it's just one almost nobody turns on by default. In Ads Manager, open Columns, choose Customize Columns, and under the Engagement section add Post Saves and Shares alongside your standard set. From there, Cost per Save is just spend divided by saves, and Cost per Share is spend divided by shares, the same way you'd already think about Cost per Click or Cost per ThruPlay.
The part that actually matters is what you do with those numbers once they're visible. Don't replace CPA with Cost per Save, that's swapping one single-metric trap for another. Instead, build a small scorecard for ToF creative specifically, one that judges a video ad on the combination of signals below rather than any single number in isolation.
| Metric | What It Tells You | Role in a ToF Scorecard |
|---|---|---|
| Cost per Save | How cheaply you're earning genuine "come back to this later" intent | Primary signal for considered-purchase categories |
| Cost per Share | Whether the ad is compelling enough to pull in a second decision-maker | Primary signal for household or joint-decision purchases |
| Thumb-stop & hold rate | Whether the hook and pacing are actually working | Confirms the creative itself is sound, separate from intent |
| Direct CPA | Same-session conversion efficiency | Secondary signal for ToF, still the primary one for BoF |
The point isn't to throw CPA out, it's to stop applying a bottom-of-funnel yardstick to a top-of-funnel job. A discount static and an educational video ad are not competing for the same conversion, they're playing different roles in the same journey, and judging both by the same number is how the video ad loses every time.
The $800 Stroller: A Case Study in Hidden Intent
At Biddyco, we recently audited an account for a premium stroller brand where this exact dynamic was playing out. A stroller is an $800 considered purchase, it's rarely an impulse buy, and the decision usually requires sign-off from a partner, which makes it close to a textbook case for exactly the blind spot described above.
When we looked strictly at direct attribution, the data told a misleading story.
| Ad Type | Objective | Direct CPA | Saves & Shares | Action Taken |
|---|---|---|---|---|
| Educational Video | Product demo (ToF) | Unprofitable / high | Massive volume | Paused prematurely |
| Static Image | 10% discount (BoF) | Highly profitable | Near zero | Scaled up |
The client had been pausing the video ads because of the high direct CPA, a completely reasonable call to make with only that number in front of you. But once we added custom columns for Ad Saves and Ad Shares, a different picture appeared: the videos were generating an enormous volume of engagement. Parents were saving them to reference later while comparing strollers, and sharing them via DM to a spouse or partner before making the call together. The videos were doing the actual selling, walking a skeptical, considered-purchase buyer through why this stroller was worth $800. The discount statics weren't generating demand at all, they were simply acting as the checkout lane once that demand already existed.
By turning the videos back on and optimizing them against Cost per Save instead of direct CPA, the account's top-of-funnel engine started running again, and overall volume and profitability recovered along with it. Nothing about the discount statics changed. What changed was giving the video ads credit for a job Ads Manager was never going to assign them on its own.
The "Silly Video" Caveat
There is one critical rule when optimizing for these metrics: the creative has to be intrinsically tied to your product. Saves and shares are a proxy for purchase intent, and a proxy is only useful when it's actually measuring the thing you care about.
If you launch a wildly entertaining meme video with no real connection to your core offer, it can rack up thousands of shares and a remarkably low Cost per Share, and still never translate into a single downstream sale. The engagement is real, but it's disconnected from the product, so it doesn't compound into anything.
| Signal | Valid Purchase-Intent Signal? |
|---|---|
| Save on a video demonstrating the product | Yes, the viewer intends to reconsider a real purchase |
| Share of a video explaining a feature or use case | Yes, they're pulling a co-decision-maker into the loop |
| Share of an unrelated meme with your logo stamped on it | No, entertainment value only, no connection to the offer |
Saves and shares are only valid indicators of purchase intent when the ad actively demonstrates, explains, or pitches the actual product. The moment the entertainment becomes the point instead of the product, the metric stops meaning anything useful, no matter how good it looks in the column.
How We Build This Into Our Process
This is part of the same account-level reporting we build for every client, not a one-off audit technique. We've written before about how our own creative process actually runs month to month, and the dashboards we use to figure out what's working start with exactly this kind of data, saves, shares, hold rate, and thumb-stop ratio, alongside direct CPA rather than instead of it. It's also the same discipline behind the account-level analysis we walked through when we broke down three Meta accounts spending $1M-$3M a month, looking past the surface-level number to understand what a piece of creative is actually doing for an account.
The stroller account isn't an outlier, it's a pattern we see constantly with considered-purchase brands specifically: anything with a real decision-making process behind it, a partner to consult, a comparison to make, a price point that requires some thought, tends to generate exactly this kind of hidden, high-save, high-share, low-direct-CPA top-of-funnel creative. If your account sells something people think about before buying, there's a real chance you have a video like this sitting in your account right now, quietly getting paused for the wrong reason.
The Bottom Line
Before you kill your next top-of-funnel ad for missing its CPA target, add Saves and Shares to your Ads Manager columns and actually look at them. You might find that your worst-performing ad on paper is the foundation the rest of your account is quietly standing on.
Biddyco is a creative agency for Meta ads built for brands spending $100k-$5M+/month who need a partner that reads an account past the surface-level number, not just a vendor optimizing for whichever metric is easiest to report.
