Puff Stand | puffstand.com
Hey - I took a look at your Google Ads account. Short version: the campaigns are doing okay, but conversion tracking is broken in a really specific way that's making every other number in the account meaningless. Once that's fixed, the rest of the cleanup is straightforward.
Across the last 30 days the account spent $1,192.64, brought in 3,800 clicks at $0.31 average CPC, and Google is reporting back "12,549 conversions worth $663,783." You did not actually sell $663k in vending machines on $1,193 of ad spend. Sorry. So the first chunk of this audit is about why those numbers exist, and the rest is about the smaller stuff that's also worth cleaning up while we're in there.
The good news: CPCs are cheap, click volume is healthy, and at a $2,700 - $5,800 machine price there's plenty of margin to make this profitable once the data is real. The bad news: right now, we can't tell what's working and what isn't, because everything Google reports back is built on a broken signal.
Hard to project numbers exactly because every dashboard metric is wrong. But here's roughly what to expect once tracking is real and the cleanup below is done:
| Right now (broken) | After cleanup | |
|---|---|---|
| Monthly spend | $1,193 | $800 - $1,000 |
| "Conversions" reported | 12,549 | 5 - 15 (real ones) |
| "Conv. value" reported | $663,783 | $15k - $40k |
| Reported ROAS | 55,627% | 1,500 - 4,000% |
| Numbers actually mean anything? | No | Yes |
The most valuable thing in this audit isn't a savings number - it's getting back to a place where you can trust what Google tells you, which is what lets you actually make decisions.
Last 30 days you spent $1,192.64 and Google is reporting back 12,548.81 conversions worth $663,782.74. Either you're making the most money any vending-machine company has ever made, or the conversion tag is broken. It's broken.
What this almost always means: the conversion event is firing on the wrong action. Probably every page view, every Add to Cart click, or a Shopify event that fires multiple times per session. The actual purchase tag should fire one time, on the order confirmation page. That's it.
This is the first thing to fix, because every other broken thing in this audit is downstream of this one. Google's bidding algorithm is using the fake "conv. value" to decide where to spend your money. So until this is right, nothing else really matters.
If you drill into each campaign, the same break shows up at the campaign level:
A conversion rate over 100% isn't a thing. Each click can convert once. So 327% means the conversion tag is firing 3+ times per click. And $0.05 conversions on a $3,000 vending machine isn't real either - that's roughly 60,000x markup. Same story as Finding 1, just sliced by campaign.
If you look at any of your search campaigns and segment by network, you'll see Search Partners and the Google Display Network are still enabled. They shouldn't be - those are garbage placements that quietly bleed search budget into low-quality syndicated sites and display banners.
Across the visible campaigns: Search Partners ate $31.42 on Leads-Search-Glamr (0 conversions), $13.04 on Canada, and $5.98 on Leads-Search-1. Not huge dollars, but it's recurring drag every month, and Search Partners traffic is consistently lower quality than the main Google Search network. Display is sitting at $0 right now but is still armed - turn it off too so it doesn't quietly switch on.
Branded queries: "puff stand" and "puff stand vending" are getting matched and counted as wins. People searching your brand name would have found you organically anyway - you don't need to pay Google for those clicks. Filter them out.
Maybe-mismatch queries: the report also shows traffic from "ID vending machine," "vending machine that verifies ID," "id-verification vending machines," "nicotine vending machine," "age verification vending machine," "Zin vending machine," "smart skincare vending," and "franchise opportunity" searches. Some of these are obviously wrong (skincare, franchise). But the brand name is "Puff" and your homepage mentions "Optional ID scanner ensures full 21+ age verification compliance" - so I genuinely don't know if you're in the age-restricted vending space (cannabis, nicotine, Zyn) or not. If you are, "nicotine" and "age verification" are your buyers, not bad traffic. Want to quickly let me know which product categories these machines are actually sold for, and I can finalize the negative list.
Inside Performance Max-1, AI Max is enabled, plus the two settings that hand Google the keys: Text customization (Google rewrites your ad copy from your site) and Final URL expansion (Google picks any landing page on your site).
Normally AI Max isn't the end of the world, but it depends on real conversion data to make smart decisions. Right now it's getting "thousands of conversions at $0.05 each on anything vending-machine-shaped," so it's widening the net into nicotine, skincare, ID-verification, etc. - because those look like wins from where it's standing.
Turn AI Max off. Turn Text customization off. Turn Final URL expansion off. Revisit all three later, after tracking is fixed and there are 30+ real conversions on the books.
Performance Max-1 is running on Maximize Conversion Value with a Target ROAS of 62,396%. That's not a typo. That's 624x return on every dollar spent.
Nobody picked that number on purpose. It's where Google's algorithm drifted to because the fake conv-value data made every bid look like it was crushing it. Even Google's own recommendations on this same screen suggest "Target ROAS 37,438%" / "49,917%" / "74,876%" - all equally meaningless, because they're calculated off the same broken value. The whole thing is a closed loop of garbage feeding garbage.
Switch the bid strategy to Maximize Clicks with a manual max CPC cap (something like $1.50 - $2.00) while tracking is being fixed. Once you have real data, then we can talk about Target ROAS.
One more small one. Inside campaign settings, the location targeting option is set to "Presence or interest" - which Google labels as "recommended" but isn't actually what you want.
"Presence or interest" basically means "people in the US, plus anyone in the world who's ever shown interest in the US" - which is a lot of people who can't buy from you. The "Presence" option below it is what you want: just people actually in the US. Google calls the first one recommended because it gets them more ad impressions, not because it gets you more sales.
Switch every campaign to "Presence." Two-second fix.
Roughly in order of priority:
The account isn't failing because you picked the wrong keywords or wrote bad ads. It's failing because the conversion tag is wrong, which has cascaded into every other bid signal Google uses. The Target ROAS of 62,396% wasn't a Puff Stand decision - it's where the algorithm drifted to because it was being fed garbage data.
The product is right. The pricing carries margin. The traffic is cheap. With tracking fixed and the noise cleaned up, this account should be profitable.
I'll go in and fix everything in this audit - rebuild the conversion tracking, build the brand and product-mismatch negative lists, turn off AI Max, switch the bid strategies, and clean up the location and network settings. Then I'll walk you (or your team) through it on a call so you can keep it running.
What you get:
$2,500 one-time
Includes 30-day and 60-day follow-ups