Wholesale Sock Deals — Brand Search Leakage

Section 01

What We Looked At

You run a dedicated brand campaign, ED | Search | Wholesale Sock Deals Brand ONLY. It exists to catch people who already know you and are searching for you by name. It does that job cheaply, at around a 9 to 10 percent cost of sale.

The question we set out to answer was simple: is anything else catching those brand searches first? When a Shopping or Performance Max campaign intercepts someone typing your name, three things happen. You pay a different, usually higher rate for a customer who was already coming to you. The intercepting campaign books the revenue, so its return looks better than it is. And your brand campaign looks smaller than the demand it actually represents.

We measured every search term containing your brand across two separate 90 day windows, then compared them. Anything catching a brand search that is not the brand campaign counts as leakage.

One important definition

Your brand name is built out of a phrase that also describes your product. That makes this analysis unusually easy to get wrong, so we drew a hard line.

Counted as brand

Searches anchored on the deals part of your name. "wholesale sock deals", "wholesalesockdeals", "sockdeals". Someone typing these wants you, specifically.

Not counted as brand

"wholesale socks" on its own, which pulled 10,633 impressions and 825 clicks. That is people shopping for a product, not looking for your store. Counting it would roughly double the headline and measure the wrong thing entirely.

We also set aside a small ambiguous group. Bare "wsd" is not safely yours. "wsd fire helmets" and "wsdm clothing" are unrelated companies. Those never enter the numbers below. Every figure in this report is deliberately conservative.

Section 02

Performance Max Is Catching A Fifth Of Your Brand Demand

Across the most recent 90 days, April 15 to July 13, here is where brand searches actually went.

Brand Clicks Missed

22.2%

of all brand searches landed somewhere other than your brand campaign

PMax Brand Clicks

580

people searching your name, caught by Performance Max

Revenue Credited

$29,982

booked by PMax off searches for your own name

Brand Campaign

$40,697

what the brand campaign itself produced, at 10.0% cost of sale

Put plainly: Performance Max is credited with roughly three quarters as much brand revenue as your actual brand campaign. These are not new customers Performance Max went out and found. They are people who typed your name into Google.

Shopping is a smaller, different problem

Your Shopping campaigns showed against brand searches 18,651 times but earned only 76 clicks, a 0.4 percent click through rate. People saw the Shopping ad, ignored it, and clicked your brand ad instead. It only cost $184.81 across 90 days, so the money is not the issue. It simply confirms brand searches are reaching Shopping, which is cheap to fix and worth tidying.

Section 03

We Checked It Twice

A single 90 day window can be distorted by one unusual month, one big order, or one seasonal spike. So before bringing you a recommendation we reran the identical analysis over the previous 90 days, January 15 to April 14. The two windows do not overlap and cover different seasons.

Measure	Apr 15 – Jul 13	Jan 15 – Apr 14	Verdict
Brand clicks landing outside the brand campaign	22.2%	19.4%	Holds
PMax brand clicks	580	396	Holds
PMax revenue credited to brand searches	$29,982	$25,851	Holds
Average order value, PMax brand searches	$628.08	$629.41	Holds
Average order value, brand campaign	$357.30	$374.44	Holds
PMax order size vs brand campaign	1.76x	1.68x	Holds

The number that convinced us

Average order value on Performance Max brand searches came back at $628.08 in one window and $629.41 in the other. Two separate 90 day periods, different seasons, measured independently, landing barely a dollar apart. That is not luck. Brand searches that land in Performance Max really do turn into bigger orders, consistently, at roughly 1.7 times your brand campaign's order size. Section 05 explains why we do not think Performance Max deserves the credit for that.

Section 04

What Is Real, And What Is Noise

This is the part we would most want you to take away, because it changes what you should and should not act on.

The totals above are stable. But the moment you ask which campaign is doing the harvesting, the numbers fall apart completely. Here is the brand revenue credited to each Performance Max campaign, in each window.

Campaign	Apr 15 – Jul 13	Jan 15 – Apr 14
PMax \| Vendor 1596	$15,304.65	$3,058.38
PMax \| Sweatpants/Joggers	$7,450.90	$0.00
PMax \| Second Zombie Vendors	$3,524.47	$0.00
PMax \| 403 \| Apparel	$799.32	$13,421.38
PMax \| Home & Kitchen	$0.00	$7,926.84

In each window, roughly half of all Performance Max brand revenue sits inside a single campaign. In April to July that campaign was Vendor 1596, with 51 percent. In January to April it was 403 | Apparel, with 52 percent. It is a different campaign each time.

What appears to be happening: a couple of large wholesale orders arrive on brand searches each quarter and land in whichever campaign happens to catch them. That campaign then looks like a brand harvesting machine, but only for that window. Note that the rotation is not even seasonal. Sweatpants and Joggers took $7,450 in the April to July window and nothing at all in January to April, which is the opposite of how sweatpants sell.

Trust the channel level

Performance Max as a whole harvests $26,000 to $30,000 of brand revenue every 90 days. That figure barely moved across two windows. It is real, and it is worth acting on.

Do not trust the campaign level

No individual campaign's brand revenue survived a second window. Any budget decision made on a single campaign's brand numbers would be a decision made on noise.

Vendor 1596, specifically

Worth separating two claims about this campaign, because only one is true. It does persistently harvest brand clicks, 157 in one window and 107 in the other, a top two catcher in both. That is structural and real. But the $15,304 it booked in April to July is a one window artifact. On comparable click volume the following quarter it produced $3,058, with average order value dropping from $646.87 to $248.71. The campaign catches your brand traffic reliably. The money it appears to make from doing so does not repeat.

Section 05

What We Think Is Happening

There are two honest readings of a 1.7x order size gap, and they point in opposite directions. It is worth being explicit about both.

Reading one: PMax is earning it

Performance Max is taking people who searched your name and successfully moving them into larger, higher value orders they would not otherwise have placed. If this is true, leave it alone. It is doing real work.

Reading two: PMax is intercepting

Large wholesale buyers search your name because they already intend to place a large order. They land in Performance Max, buy what they came for, and PMax books it. If this is true, you are paying vendor campaign rates for sales your brand campaign would have closed at around 10 percent.

The rotation is what tips it

The second reading fits the evidence better, and the reason is section 04. If Performance Max were genuinely upselling brand searchers, the campaigns doing the upselling should be reasonably consistent from quarter to quarter. They would be the ones holding the right inventory for the job. Instead, a rotating cast of unrelated campaigns catches a stable pot of money. Vendor 1596, then 403 | Apparel. Sweatpants one quarter, Home and Kitchen the next.

That pattern is the signature of demand that already exists and simply lands wherever it lands. On that reading, the order size gap is telling you something real but different from what it first appears: it reflects who searches for you by name, which is disproportionately your bulk buyers, rather than anything Performance Max did to them.

The order size gap is real. The claim that Performance Max created it is the part that does not survive contact with a second window.

We want to be straight with you: this is an inference, not proof. Aggregate campaign data cannot fully separate "Performance Max upsells brand searchers" from "bulk buyers happen to land in Performance Max". What has changed is where the burden sits. The upsell explanation now has to account for why the upselling campaign is a different one every quarter. We do not think it can.

Section 06

What We Recommend

Rather than argue the interpretation, we would rather settle it. There is a clean test, and it costs nothing but patience.

Run the brand exclusion test on Performance Max. Exclude brand searches from Performance Max and watch what happens to total brand revenue across the account, not to any one campaign. If that revenue reappears in the brand campaign at around 10 percent cost of sale, it was interception and you have just cut your cost on the same sales. If total brand revenue drops, Performance Max was genuinely doing work and we put the exclusions straight back.
Give the test at least 90 days. This is the part we want to be firm about. Given that one or two large orders drive half the quarterly total, a 30 day read will either catch a big order or miss one, and either way it will tell you nothing. A short test here is worse than no test, because it produces a confident answer that is wrong.
Add brand negatives to the Shopping campaigns now. This one needs no debate. It is only $184.81 across 90 days, but it is trivial to do and it cleans up the reporting so brand demand stops muddying your Shopping numbers.
Stop reading brand revenue at the campaign level. Immediate, free, and prevents a bad decision. When you look at Vendor 1596 or any vendor campaign's return, understand that some portion is brand demand it did not create, and that the portion is not stable enough to quantify per campaign. Strip brand at the channel level instead.

What we are not recommending

We are not recommending you cut or restructure Vendor 1596 on the strength of its brand numbers, which is where this analysis pointed before we ran the second window. That campaign's $15,304 looked like a strong reason to act. It was a single quarter's luck. This is exactly the kind of decision the second window was run to prevent.

Section 07

What This Cannot Tell You

Three limits worth knowing, so nobody builds a decision on a number that will not hold it.

We cannot price the leak

Google does not report cost for Performance Max search terms. Impressions, clicks, conversions and revenue, but no spend. So we can tell you 580 brand clicks went to PMax, but not what you paid for them. Any spend figure here covers Search and Shopping only and is a floor, not a total.

PMax volume is a minimum

Google withholds an uncategorised bucket inside Performance Max search data that cannot be inspected. The real brand volume reaching Performance Max is at least what we have reported here, and possibly more.

We cannot prove intent

Campaign level data shows what happened, not why. Settling reading one against reading two for certain would need order level data from the store, and a median order value rather than an average. The brand exclusion test sidesteps this by measuring the outcome directly.

How to reproduce this

Both windows were pulled straight from the Google Ads API with the same scripts and the same brand definition, and everything is version controlled on our side. Any figure in this report can be traced back to its raw pull on request, and the analysis can be rerun over any window you want.

Where Your BrandSearches Are Landing