The #1 reason your Demand Gen isnโt scaling?
๐ Youโre optimising for last click.
Start scaling on first click performance and watch what happens.
๐ Want more scaling insights? Follow me for proven Google Ads strategies.
#GoogleAds#DemandGen
Google Ads has a weird way of making simple things unnecessarily complicated.
Demand Gen literally gives you a separate option to create video ads.
Yet when setting up an image ad, Google also gives you a setting to automatically create video assets from those images.
Yes, you can turn that setting off.
But my question is: why does this setting need to exist inside an image ad setup in the first place?
If I select an image ad, the default expectation should be:
๐ Image assets = image ad
๐ Video assets = video ad
If I want video, Iโll create a video ad.
I understand the reason Google is doing this โ inventory coverage and not performance. But from a media buyerโs perspective, it makes the creative format unnecessarily blurry.
Especially when testing creatives.
If Iโm testing a static image, I want to know:
โHow does this image perform?โ
Not:
โHow does Googleโs automatically generated video version of this image perform?โ
Those are two different creative tests.
The option to disable it is there, but honestly, I donโt think the option should be there at all.
Image ads should be image ads.
If I want to give Google permission to create video assets for additional inventory, that should be something I explicitly opt into โ not something that needs a toggle inside an image-ad setup.
"New Google Ads accounts need to warm up slowly."
Not really.
Took one live last week. Completely fresh, no history, no conversion data.
It's at $1.3k/day now, about 80% of that in Demand Gen.
If it holds, $8-10k/day within the month.
Account age isn't what holds these things back. It's everything you set up before launch.
$15K+ day at 1.8x ROAS
This came from image ads > Demand Gen Only
A lot of accounts still treat Demand Gen like a side campaign, but if the creative is right, it can drive serious volume
Still room to push from here
This is the kind of scale you can expect when you work with us.
If you want someone who can genuinely scale your account and push hard this Q4, DM me. ๐
Most brands treat Google as the channel that closes people Meta warmed up.
Then they wonder why growth stalls.
This supplement brand did 1.52 ROAS on first-time buyers alone last month. $39.5K profit on $30.4K spend.
New customers. Acquired profitably. No burning cash to "buy growth."
Google isn't your closer. Set up right, it's your acquisition channel.
Most GMC suspensions happen before you spend a dollar.
The site is the problem. Missing return policy. No contact page. PDP claims Google hates.
So I built a skill that audits your entire site for suspension triggers BEFORE you create Merchant Center.
Full audit. Every flag. Action items to fix each one.
Want it?
Like this post + comment "AUDIT" and I'll DM it to you (must be following)
People ask me how to capture competitor traffic all the time.
Most run a standard competitor Search campaign, watch QS tank, pay stupid CPCs, and give up.
Here's what works instead:
Asset-only PMax. No feed.
1. Competitor names + variants as search themes
2. Custom segment audience built from their website URLs
PMax matches to competitor demand without the QS penalty crushing you.
Way better efficiency than the Search version everyone tries first.
Found a pumpkin seed oil brand running one of the cleanest advertorial.
The structure:
Opens with a character. Diane, 54, office manager, Georgia.
Not the product. A person the reader recognizes.
Then 7 symptoms.
Reader self-diagnoses before the pitch even starts.
Then it names a villain: DHT.
Not "aging." A hormone you can BLOCK.
Then it kills every alternative.
Shampoos, biotin, kegels. All dead before the product shows up.
Then a real 24-week study.
The 40% number does the selling, not the copy.
The product isn't even named until you're halfway down the page.
Cold traffic doesn't buy from a PDP on first visit.
It buys the story that explains why nothing else worked.