Oct 27, 2025
Will this YouTube ads strategy actually work?
We ran one of Germanys biggest YouTube ad agencies and have spent well over $10 million on YouTube.
Meetings with Google ads officials to develop our strategy
Clients getting million-dollar-awards with this strategy
Our seven $1M awards and our $10M award earned with this YouTube ads strategy
So it's safe to say this strategy is well-tested and proven to work.
Advantages of YouTube ads
The ad performance is extremely stable
We've had YouTube ads that ran profitably for over 1.5 years without any changes
So YouTube ads can be a lot more "set and forget", compared to let's say Meta ads, where you have to refresh your ad creative every few weeks
You reach audiences with clear buying intent
For example, a person searches “How do I train my dog?” on Google or YouTube
They’ll then get YouTube ads about products like a dog training video course (perfect match for their intent)
On other advertising platforms like for example Meta, they have zero intent. They’re just mindlessly scrolling
Awesome targeting options
You can target people that performed certain Google or YouTube searches, people that visited certain websites, people that watched your organic YouTube videos or subscribed to your channel, etc.
You often get free ad credits from Google to get started
At the moment you get $400 in free ad spend after spending $400 of your own money
So you basically get a 50% discount on ad spend to get started
Google is a very robust ad platform
Good advertiser support
Very few ad manager bugs
Very few ad delivery outages
Very few ad account bans
Very few ad disapprovals
Who do YouTube ads work best for?
Almost every offer/niche/product can work on YouTube, but generally speaking, the more mass market your offer is, the better.
Setting expectations
At the beginning, YouTube ads can definitely be hard to crack. But once you've found your winning ad messages and winning audiences, there’s no other ad platform with such stable and smooth ad performance and with such huge scaling potential. Once you crack YouTube ads, not even the sky is the limit. Okay, maybe I’m exaggerating a bit, but you get the point. So be patient and have faith in yourself. It’s 100% worth it.
Step 1: Account setup
Create an Google Ads Manager Account (MCC)
A “Manager Account” houses multiple different ad accounts
Create as many Ad Accounts as you’re allowed in your Manager Account
We do this in order to have multiple backup Ad Accounts in case one gets banned
If you already have existing Ad Accounts, connect them with your Manager Account
You could also use third-party tracking softwares (optional)
Create a YouTube channel (if you don’t have one already)
Step 2: Creating ads
With YouTube ads the ad creative is the most important part
The “Ad Creative” consists out of the “Ad Visuals” and the “Ad Message”
Ad Visuals (Only 10% - 20% importance)
Everything the user sees (B-roll, presenter, outfit, room/background, …)
This is for building trust, credibility, and rapport
Example: Having a professional-looking doctor present your ad message for a blood sugar offer will perform a lot better than having a sleek businessman performing that ad script
Because people trust doctors with their health and won’t trust businessman with their health
The visuals are there to amplify and boost the ad message
Ad Message (80% - 90% importance)
The words that are being said, so everything the users hears
If the ad message/ad script is bad, nothing else will be able to save the ad
The ad message has the following variables:
How you structure the ad
Unaware audience: Create a problem -> Twist the knife -> Offer solution -> CTA
Problem-aware audience: Remind them of the problem -> Twist the knife -> Offer solution -> CTA
Solution-aware audience: Differentiate your solution from other solutions -> Prove your solution is the best -> CTA
Product-aware audience: Explain your solution -> Handle all objections -> CTA
Most aware audience: Give them a good reason to buy right now -> CTA
Having the right structure will make your ads work extremely well, and having the wrong structure will lead to poor performance
The language/slang/jargon you’re using
If you’re using formal language when trying to advertise to teenagers, it won’t work
But using formal language and technical terms when trying to advertise to lawyers or doctors will work really well
You have to sound, feel, and appear like you’re one of them in order for the ads to work. If you don’t, they won’t work. You have to have rapport with the audience.
How you’re hooking your target audience
The goal is to grab only the attention of your target audience and completely ignore everyone else
If the hook is really specific and grabs the attention of only your target audience, the ads will work really well, and if the hook is too broad and grabs everyone’s attention or is so generic it grabs no one’s attention, the ads won’t work.
The problems you’re addressing
Your target audience is most likely experiencing multiple problems related to your offer
If you’re helping overweight people to lose weight, they’re not only experiencing the core pain of being overweight but also other problems that derive off that, like not being able to go to the beach without everyone staring, not being able to find trendy clothes that fit, not being able to play with their kids without being out of breath, being a bad role model for their kids, and so on…
And some problems will make your ads convert better than others
So using the right problems will make your ads convert really well, and using the wrong problems will make your ads perform really weakly.
How you frame your solution
How you frame or position your solution/offer effects conversion rates
You could frame a weight loss offer as a “Weight loss coaching” (They’ve heard that a million times), or you could frame it as a “60-Pounds-Takedown-Challenge” (New, shiny, interesting)
Framing your offer the right way will make your ads work extremely well, and framing your offer the wrong way will make your ads fail
How you frame the call-to-action
How you frame or position the next step in the funnel also effects conversion rates
You could frame a VSL or Webinar as a “Free video presentation” (Boring), or as a “Step-by-step training video” (Slightly better), or as a “Video where you’ll break down your entire process, pull back the curtain, and give exclusive behind-the-scenes looks into real case studies and their results” (New, shiny, interesting)
Framing your CTA the right way will increase conversions a lot, and framing the CTA the wrong way will absolutely kill conversions
How you tie it all together into an congruent sales argument
The entire argument has to follow logical steps
It doesn’t make sense to argument that “Amazon FBA is the best way to make money online” if the person doesn’t even know what “Amazon FBA” is (You skipped a logical step in the argument)
And it doesn’t make sense to argument that “Remote closing is the best way to quit your 9-to-5 job and work from home” if the person first hasn’t been convinced that having a 9-to-5 job is bad and therefore desires to work from home (You skipped a logical step in the argument)
Having a flawless, congruent, and logical sales argument will make your ads take off, and having a mismatched sales argument will make your ads fail
As you can see, there are a lot of variables that you have to get right in order create a winning ad message from scratch
But luckily there are two shortcuts for creating winning ad messages/ winning ad scripts
Shortcut 1: Model winning YouTube ads that are already proven to work
You can use the Google ads transparency center, but sadly you can’t see any relevant data, you can’t identify the winners, and you can’t search for keywords, only for specific advertisers or websites. So it’s hard to discover new ads.
We always use ytads.com for that. It’s possibly the biggest YouTube ad library with also the most data per ad. You can find millions of winning ads, spy on your competitors' ads, and see data like views, ad spend, landing page URL’s, word-for-word transcripts, etc.
I also created a swipe file of the 349 highest-converting YouTube ads I found so far. There are ads for all kinds of different offers in there. Maybe you’ll find some winning ad messages to adapt for your business.
Shortcut 2: Let AI write the ad message/ ad script for you
But please DO NOT use some random AI models and expect the ad scripts to work. Regular AI models don’t have any real performance data about YouTube ads to go off. All you’re doing is letting an AI guess what ad message could potentially work, instead of guessing yourself.
You could gather thousands of YouTube ad scripts, including their actual performance data, and give an AI model that data. Now the AI knows what ad messages actually perform in the real world and can reverse-engineer those for you.
But we use YTADS for that as well. The AI has access to the entire database of millions of winning YouTube ads with data from over $6 billion of ad spend. It then uses pattern recognition to reverse-engineer the winning ads and write your own pre-tested ad script based on the proven ads from the library.
At the beginning, when you have no winning ads yet, test at least 2 new ad creatives a day
Test wildly different ad messages until you find the general direction that’s working. We need big swings here. Small variations of the same ad message are useless at this stage.
And since creating each video ad manually is pretty time-consuming, we also use a shortcut for that
Instead of creating every single ad manually (which would take way too long), we test new ad messages rapidly by using AI-generated YouTube ads
Let’s say we want to test 20 new ad messages
We generated those 20 YouTube ads in about an hour using AI
We launch them and let them gather some data (for about 3 - 5 days)
We evaluate their performance
We pause all AI ads
We then re-film the ad messages of the winning AI ads authentically using a real human. Because now we can actually justify that effort because the ad messages are already proven to work in our ad account for our offer.
This way we get the best of both worlds. We can rapidly test new ad concepts using the AI-generated ads. But then later also have authentic versions of those YouTube ads to scale with.
But of course you could also scale the winning AI ads without re-filming them. Well-made AI ads definitely work. I’ve seen people make millions with completely AI-generated YouTube ads.
You can create AI YouTube ads by hacking a bunch of different tools together in your own automation workflow
But we use YTADS to generate the AI YouTube ads. It’s just a few clicks and takes about 6 minutes per ad. It turns the ad script into a voiceover, adds a lip-synced presenter, and adds background music, b-roll footage, sound effects, zooms, transitions, etc., all for you.
Here are some examples of those AI-generated YouTube ads
Once you’ve found some winning ads, don’t stop testing new ads!
Continue testing at least 5-7 new ads a week and start to dig deeper into the ad angles that are already working
Because in case your winning ads should suddenly stop working, you need to have new winning ads to replace those
Ads are never static; they’re always moving, rotating, increasing & declining
This is what will happen if you stop regularly testing new ads:
You won’t find new winners...
Sooner or later, your current winners will decline…
You won’t have new winners to replace them…
You’re ad account will dry up, and everything will stand still…
Now you’re back at zero...
So don’t let that happen in the first place, and continue to test new ads even though you’ve already got some winners!
Step 3: Targeting
Start with the easiest audience type on the top and then only move to the next harder audience type once you’ve got the previous one to convert profitably
We want to start with the least resistance and work our way towards the audiences with the most resistance
Because if you can’t get the easiest targeting and the hottest audience to convert, you certainly won’t be able to crack the harder and colder ones
But you don’t have to use or test every single one of those targetings. Only test the ones that make sense for your business or offer
If, for example, Google’s “life events” segments don’t make sense for your offer, you could definitely skip those
Or if for example, you don’t have customer list to upload for the “customer match” targeting, you could skip that as well
And once you’ve found your first 1-2 targetings that convert, you can basically stop testing targetings for a while and ride those winning targetings until you’ve absolutely capped out those audiences
And only once you’ve hit the scaling ceiling with those audiences, only then you move on to the next harder ones
So don’t move to the next harder targetings too soon. Only once the easier winning targetings have hit diminishing returns
We don’t want to make it hard on purpose. So if you’ve found a winning targeting, scale it as much as possible before moving into the next colder audience
And if you move to a larger and colder audience, you will most likely have to tweak your entire sales message...
You will have to tweak the messaging of your ads to resonate with the colder audience and the new market awareness stage...
You will have to tweak the messaging of your funnel to match with the new messaging of the ads…
And you will have to tweak your offer to make it more “cold friendly”
The offer has to be even easier to understand, so it has to sound even less complicated...
It has to solve an even more widespread problem...
It has to sound and feel even more “sexy” and new…
It has to get results even quicker...
It has to involve even less risk...
It has to feel easily attainable, ideally with lots of proof…
And the larger and colder the audience gets, it gets even more important to test a lot of new ad messages and a bunch of unique ad angles to see what works and what doesn’t
With a small, tight audience, it’s fairly simple to predict what they want exactly
But the bigger the audience gets and the more widespread it is, the harder it is to find out what to say in order to resonate with all of them at once
Here you don’t start by testing small variations, but rather big swings to see in which general direction the messaging has to fall
You have to look at your offer from different perspectives
They might resonate with “saving money”
Or maybe they resonate more with “making more money”
Or maybe they resonate more with “investing their money”
Or maybe they resonate more with “best side hustle”
Or maybe they resonate more with “best online business to start”
Or maybe they resonate more with “second income next to your job”
Or maybe they resonate more with “how to start a business to quit your job”
Or maybe they resonate more with “start a side hustle to support your family”
Or maybe they resonate more with “reaching financial freedom”
You won’t know what angle connects best until you test it
And once you’ve found that winning ad angle, then you can double down on it and create smaller variations of that to fine-tune it
Step 4: Launching ads
Campaign structure: 1 Campaign -> 1 Ad Group (containing 1 targeting) -> 3 - 5 ads
The campaigns are separated by the targeting
The ad groups contain multiple ads to test more quickly and more efficiently. The alternative would be “1 Campaign -> 1 Ad Group -> 1 Ad”, which works great too but is a lot more work to set up.
Don’t put too many ads in one ad group; otherwise, the budget gets spread too thin
Here is an example on how it would look if you were to test 3 new ads with 3 different audiences
How to set up a YouTube ad campaign:
Open your Google Management Account and select the Ad Account you want to use
In the dashboard, click on “New campaign”
Select your campaign objective. If you’re going for conversions, sales, and driving revenue, select the “Sales” objective
Add the conversion events you want the campaign to optimize for (e.g., opt-ins, scheduled calls, checkouts, sales, etc.) and then click on “Continue”
Then select the “Demand Gen” campaign type and click on “Continue”
Because the “Video Action” campaign type on the bottom of the image isn't available anymore. Clicking it will also create a "Demand Gen" campaign.
But “Demand Gen” campaigns are actually even better
Give your campaign a name
Make sure to use a consistent naming scheme across your entire account
Otherwise, it’ll get messy really quick
Either create your own naming scheme that works best for you
Or use this: “Offer name - Targeting name - Creative batch numbers”
E.g., “30-Day Abs Challenge - INTENT: Competitor Websites - Ads #1 - #3”
Select either “Conversions” or “Conversion Value” as the campaign goal
“Conversion Value” only works if you’ve actually set different values for what each conversion is worth to you
E.g., for you, a “front-end purchase” might always be worth $7, while an "Upsell #1 purchase” might be worth $297
In this case, if you’d select “Conversion Value”, Google would focus more on the ads that bring in a lot of upsells because they’re worth a lot more, rather than focusing on the ads that mainly bring in just front-end purchases
If your conversions don’t have a fixed value, like, for example, “Opt-ins” or “Scheduled calls”, you could either break it down and calculate an average value
E.g., if every 1,000th opt-in/lead buys a $10,000 coaching from you, on average every opt-in would be worth $10 to you
E.g., if every 10th scheduled call results in a $10,000 coaching sale, on average every scheduled call would be worth $1,000 to you
Or you could simply select the “Conversions” goal, and Google will simply focus on the ads that bring in the most conversion overall
Select the conversions the campaign should be optimizing for
Decide between optimizing for “Maximize Conversion” or “Target CPA”
If you don't have any previous conversion in this Ad Account for the offer and funnel you’re wanting promote with the ads, don’t check the box for “Target CPA”, then Google will use the “Maximize Conversions” optimization
In this case, Google will use your budget to simply get as many conversions as possible
If you’ve already got at least 50 conversions for this offer and funnel in your Ad Account, then you could try checking the box and using “Target CPA”
In this case, set the T-CPA (Target Cost-Per-Acquisition) to your break-even cost-per-acquisition and add 20% to that
E.g., if your front-end product costs $7, your break-even CPA would be $7, and by adding 20% on top of that, your Target CPA would be $8.4.
And this doesn’t mean Google will always spend $8.4 to acquire a customer, because then we’d be losing money. If Google is able to acquire customers cheaper, it will.
And by adding a 20% cushion, Google has a bit more breathing room and more freedom to work with
Most of the times this is a good starting point, but maybe you need to experiment with this a bit yourself
And the reason for switching to T-CPA is because often it just scales better
Set your daily budget and make the campaign start on the next day
We want the campaign to start a midnight right at the beginning of the next day
Because when we start a new campaign in the middle of the day, oftentimes Google is really inefficient with the budget and quickly spends it all before the day is over
And if we count that up to a year, that inefficiently spent budget racks up
But if we start the ads at midnight, Google has a whole day to strategically spend the budget and doesn’t have to rush things
If you’re selling low-ticket items or low-ticket offers, you could set the daily budget to 1x or 2x your average order value
And if you’re selling high-ticket offers, you could set the daily budget to 1x or 2x your front-end conversion cost
E.g., if your high-ticket funnel starts with an opt-in and if you know that one opt-in/lead costs about $20 via ads, then you could either start with $20 a day or $40 a day
But you could also just set the budget to whatever you feel comfortable with; there are no hard rules here
Just know that the lower the daily budget is, the longer you have to wait to get data and results
So a higher daily budget versus a lower daily budget basically makes no real difference; with a lower budget, you just have to wait longer
Your goal should be to get at least 1-2 conversions a day; otherwise, it takes way too long to evaluate what ads work and what ads don’t
In the devices, exclude “TV Screens”
People that watch YouTube on a TV are typically not in the mood to buy stuff; often they just want to relax on their couch after a hard day of work
With YouTube users on a desktop or mobile phone, that’s different; they are often actively search for solutions to their problems and are willing to go through sales funnels to buy stuff
On top of that, on most TVs it’s hard to click on the ads, even if you wanted to. I mean, would you know how to click on a YouTube ad with a TV remote? I certainly don’t.
So by removing TV's, we can minimize some inefficient/wasteful ad spend
Most of you will be able to ignore these settings and just leave them default
Disable the “Google Video Partners”
Often these sites and apps only bring in junk traffic
E.g., kids clicking on ads to get rewards in mobile games
E.g., people accidentally clicking on ads that suddenly pop up on websites
Click on “Go to Ad Group”
The Ad Group is where we’ll set the targeting
Name your Ad Group
Again, use a consistent naming scheme across your whole account
If you’ve used “Offer name - Targeting name - Creative batch numbers” for your campaign, then simply use the “Targeting name” as your Ad Group name
E.g., Campaign: “30-Day Abs Challenge - INTENT: Competitor Websites - Ads #1 - #3”
E.g., Ad Group: “INTENT: Competitor Websites”
Select the locations that make sense for your offer
Try not to be too restrictive
Oftentimes, the smaller the total audience gets, the higher the CPM will be because there’s less ad inventory
Select the language that makes sense for your offer
Since we want to run specifically YouTube ads, we only select the YouTube placements
Select the YouTube placements that you want to run
“In-stream” and “in-feed” are both horizontal ads
“Shorts” are vertical ads
Either select an audience you’ve already created or create a new one
Only use one targeting-type per audience
So don’t mix things up; we want to keep everything separated and only have one variable per audience
E.g., don’t layer custom intent segments, affinity segments and in-market segments all in one audience
Let’s say this audience performed great; now we wouldn’t know which of these 3 different targeting-types was the one that actually drove most of the results
But if we only have one targeting-type per audience and that audience performed great, we know exactly that this one targeting-type was responsible for that
When creating a new audience, use the same “Targeting name” from your Ad Group naming scheme as the audience name
Pick and set up only one of the targeting types
If you want, you can add exclusions
But often that’s not really necessary
Set up the demographics
Only restrict the demographics if you’re 100% sure
Because restricting the audience too much can increase CPM and make it harder to get conversions
If you’re unsure, oftentimes it’s better to leave it more broad
E.g., when selling woman shavers, you might think it would make sense to exclude all men. But maybe the men would’ve bought the shavers as gifts for their girlfriends
If you’re experiencing low-quality traffic, it may help to exclude all “unkowns”, but at the beginning I’d include them
Turn off the “Optimized targeting”
Usually this does not help and only makes it less transparent on who the ads are actually being delivered to
Later you could split-test whether this improves results in your account or not, but at the beginning simply turn it off
Unless you’re using some third-party tracking setups, most of you can ignore this
Then click on “Go to Ad”
Select “Video Ad”
It’s also possible to run image ads on YouTube
But those don’t perform as well as video ads
Video ads are the “real” YouTube ads
Then give your ad a name
Ideally, you have a folder where you’re collecting all of your ads in one place and numbering them
Then use the same creative number from that folder for the name of the ad so it matches up
E.g., Name in your folder: “Ad #1”, Name in Google ads: “Ad #1”
Enter in your funnel/landing page URL
Make sure your funnel is working properly both on desktop and mobile and nothing is broken
Paste in the link to your unlisted YouTube ad
If you want to run normal horizontal YouTube ads as well as vertical YouTube short ads, upload two ads
One horizontal version of the ad and one vertical version of the ad
But this has to be the same ad; the only difference is the aspect ratio
So ideally the ad script, the ad message and the video itself should stay the same
But sometimes that’s not possible since YouTube Short Ads have to be less than 60 sec. long
In that case, just try to keep the Short Ad as close and as similar as possible
This is not where we put in all the different ads with the different ad messages we’re testing
This is within one ad
Click on “Choose where your videos show” and then click on “All formats”
Set all horizontal ads to “in-feed” and “in-stream”, excluding “Shorts”
Set all vertical ads to “Shorts”, excluding “in-feed” and “in-stream”
This way we can make sure we’ll always have the optimal aspect ratio for the given placements
Because we don’t want to run our vertical video ads in the horizontal placements, and we also don’t want to run our horizontal video ads in the vertical placements
Next, add your logo
Turn off the “Video enhancements”
This does not work well and typically messes up the ads
If you want to run normal horizontal YouTube ads as well as YouTube short ads, don’t let Google adjust the ad for you
They will crop it randomly and the most important parts of the ad might get cut off
Ideally, you crop and adjust the ads yourself or create separate ads for in-stream + in-feed ads (both horizontal) and short ads (vertical) from the start
Enter in some copy in the text fields
Always check in the preview on the right where the text is being displayed in the different ad formats
Sometimes it’s a bit confusing which text goes where
I would not use the “Automated” Call-to-Action, since then we don’t have control over what will be displayed
And whether you should use the “Sitelinks” depends on your business and sales process
If you only have one landing page that all of the traffic should be sent to, then the sitelinks won’t make sense
But if for example, you’re an e-commerce brand and there are a lot of different sub-pages in your online shop that you could send the traffic to, then it would make sense and you could create a sitelink for each of the different sub-pages
Because then your ad will take up more space on the screen, which means the ad gets noticed easier, and the users can pick a page that’s most relevant to them, which can increase the conversion rate
Unless you’re using some third-party tracking setups, most of you can ignore this
What we’ve built so far is this
1 Campaign -> 1 Ad Group (with 1 Audience/Targeting) -> 1 Ad
So if we want to test 3 different ads inside this campaign, we click on “Duplicate ad”
Then we change the name of the duplicate ad to “Ad #2”
Remove the previous “Ad #1” and paste in the unlisted YouTube link to the “Ad #2”
Make sure to set the formats for the videos again
The rest of the settings can stay the same
Now our campaign is looking like this
1 Campaign -> 1 Ad Group (with 1 Audience/Targeting) -> Ads #1 - #2
So duplicate the ad one more time and change the ad to the “Ad #3”
Now our campaign is looking like this
1 Campaign -> 1 Ad Group (with 1 Audience/Targeting) -> Ads #1 - #3
The campaign structure in the sidebar should look like this
And if we want to test these ads #1 - #3 to two additional audiences/targetings, we would do the following:
Click on “Go to review”, double-check all the settings and then click on “Publish campaign”
In the sidebar, click on “Campaigns” to get to the campaign overview page
Select the campaign you’ve just created, click on "Edit”, and then click on “Copy”
Then deselect the campaign
Then click on “More” and on “Paste”
Then click on “Paste” again
Now we’ve duplicated that campaign
To see all campaigns, make sure you click on “All campaigns”
Now click on the duplicated campaign
In the Ad Group, click on the “Settings” icon
And then remove the duplicated audience and add another audience/targeting you would like to test
Afterwards, click on “Save all changes”
And change the naming schemes of the duplicated Campaign and the duplicated Ad Group to fit the new audience/targeting that you’ve set up by clicking on the “Edit” icon
Now we’ve built this
So simply do the process of duplicating and changing a campaign one more time
Now we’re done; we’ve successfully launched 3 ads to 3 different audiences/targetings
Depending on your experience, this might seem a bit complicated at the beginning
But this is actually fairly simple, and after a while you’ll get used to this and will be able to do this really quickly (in about 10 minutes)
Now, don’t touch the ads for at least 3–5 days
We have to let the ads gather data before we can evaluate their performance
You can check them, but don’t touch them and don’t change anything
Step 5: Optimizing
Optimizing the ads is pretty straightforward
But remember, only start optimizing once the ads have ran for at least 3–5 days
First, make sure to customize your columns to show you all the data you need
Then look at the different campaigns and dig into the ad-level
Analyse the ads and pause all ads that are performing far outside the target KPIs
Make notes on what ads have failed and try to set up some hypotheses on why the ads could have failed
Also, make notes on what ads are performing great and again, try to set up some hypotheses on why they could have performed great
Now you’ve trimmed the fat off the campaigns and the campaigns are now probably left with less active ads than before
That’s why we now want to fill them up with new ads
Try creating and adding new ads that are either really similar to the current winners
Or try creating and adding ads with new concepts that are drastically different from the current losers
You can add new ads to a campaign by clicking on the campaign, going to the ad level, and then clicking on the “Plus” icon
By adding new ads into the campaigns, we want to challenge the current winners in the campaigns
So want to see if we can create ads that outperform the current winners
If you see that a campaign is performing well and heading in the right direction, you can double the budget every 3–5 days till you’re at $500 a day
Above $500 a day, I’d increase the budget by 10%–20% every 3–5 days, as long as the campaign is still getting steady results
If you’ve increased the budget too quickly and the results start going downhill, you can also decrease the budget again to the previous level, and often the results will stabilize again
After you’ve reached 50 conversion, you can also try duplicating your campaigns and changing them from “Maximize conversions” to “Target CPA” (as described earlier in this document) to see if it improves your conversions
If you’ve got a winning audience that contains multiple keywords, multiple websites, multiple search terms, etc., you can also try splitting them up and creating separate campaigns with each of the best performing ones
You can also use the winning ads that you’ve found to test new audiences
Ideally, test new audiences that are as similar as possible to the current winning audiences that you’ve found
Every new campaign that you’re creating in this step has to go through the whole process again
So first you launch the campaigns
Then you wait 3–5 days for them to gather data
Then you analyze and optimize them
And don’t forget the by far most important step: Regularly creating and testing new ads with new ad messages and new ad angles
You have to have a constant flow of new ad creatives being launched in your ad account every week
And don’t get discouraged; the very first ads you launch rarely hit it out of the park
You have to let your Ad Account warm up a bit and give the algorithm some time to gather and process the data
Because if your Ad Account is completely fresh, the algorithm doesn’t have any data to go off yet and won’t be able to perform as well as a seasoned Ad Account with tons of history conversion data
But if you stay consistent, you’ll find your winning ads sooner than you think
Step 6: Scaling
Scaling your winning campaigns is pretty straight-forward as well
If the campaigns are performing stable within our target KPI’s, every 3–5 days we can increase the daily budget of the campaigns as explained previously (doubling till $500 a day, then 10%–20% increases)
If we’re using Target-CPA Campaigns, over time we can gradually increase the Target-CPA value
Then Google is able to spend more per acquisition and we can scale more aggressively by “outbidding” all competitors
And our goal is to maximize the ad spend per audience/ targeting
Because every audience/targeting is capped. You can’t scale them infinitely
There will be a point where you hit diminishing returns and an increase in daily budget won’t result in an increase of conversion
That’s when you hit the maximum daily ad spend that audience/targeting is able to handle
You could also call that the “audience-budget-sweetspot”
So if you’re increasing the budget of a campaign and you notice that instead of increasing or staying the same, the performance always starts decreasing after a certain amount of daily ad spend, then decrease the ad spend back to the previous amount that has performed great and ride the “audience-budget-sweetspot” that you’ve just found like a wave
And if you’ve hit the sweetspot for one audience/targeting, you’d scale by reaching that sweetspot for even more audiences/ targetings
And if you’re scaling your campaigns and digging into colder and colder audiences, one of the most important things is to adapt the messaging of the whole sales process (ads, landing page, VSL’s, sales calls, offers, etc.) to fit the new market awareness stage of that audience in order to resonate with them
This is explained in detail in the “Step 3: Targeting” section of this post
Winning YouTube ads swipe file
Here’s our swipe file of the 349 highest-converting YouTube ads. Use those ads as inspiration. Maybe you’ll find some winning ad messages to adapt for your own offer.
And if you found this post helpful, please consider sharing it with someone that could benefit from it as well :)
![](data:image/svg+xml,<svg display="block" role="presentation" viewBox="0 0 415 29" xmlns="http://www.w3.org/2000/svg"><path d="M 0 5.754 L 204.834 2.254 L 408.829 0 L 279.5 7.5 L 145 13 L 243 17.5 L 216.5 21.5 L 205 23.5" fill="transparent" height="23.5px" id="wxlBUhIda" stroke-dasharray="0" stroke-linecap="round" stroke-linejoin="round" stroke-width="4" stroke="rgb(230, 55, 82)" transform="translate(3 3)" width="408.8293105604056px"/></svg>)
