top of page

The Modern Marketing Funnel: How to Use AI to Get More from Your Ad Spend

Which KPIs are you optimizing for in each layer of your marketing funnel?

With the rapid pace of ad platform algorithm improvements, brands can squeeze a lot more out of each layer of their paid media funnel simply by choosing the right KPI to focus on, so it’s worth continuously reevaluating your methods.

In this article, I’ll provide examples of tests that you can run in your paid media campaigns and share tactics that I’ve seen improve results for my ad accounts this year.

To kick things off, let’s look at a simple 3-layer paid media marketing funnel:

Each layer of the funnel presents opportunities to test and make data-driven decisions.

With the right formula for each funnel layer, you can maximize your ad spend and scale more efficiently by leveraging the machine learning that the popular ad platforms offer.

No data science degree required - here’s an inspiring quote by Cassie Kozyrkov:

“A common mistake businesses make is

to assume machine learning is magic, so

it’s OK to skip thinking about what it

means to do the task well” - Cassie

Kozyrkov | Chief Decision Scientist, Google

To start, let’s break down the top-of funnel awareness layer.


For many brands, their awareness campaigns may be measured by impressions, CPM, reach or traffic. While all of those metrics hold value, there may be creative ways to leverage the advanced machine learning of the ad platforms to ensure you’re spending your paid media budget in the best way to drive meaningful results.

For example, instead of launching a basic reach or traffic campaign that shows your ads to users with a limited amount of data (e.g. the algorithm may only have surface level data such as whether they fit your target audience or whether they clicked on your ad) you can set up a feedback loop with slightly further down-funnel metrics such as scroll depth or actual web page engagement, then train the algorithm to not only find users within your target market, but find users that are also showing intent by measuring them one step further.

The machine learning that the ad platforms have built out is there to empower your brand, you just have to provide it clearly defined goals so it can help you create awareness with the right segment of users.

This may sound like I’m recommending that you omit the awareness phase, but I’m not. What I’m recommending is to think critically about how you’re defining your target audience and measuring awareness, because in today’s society of short attention spans, just gaining a quick impression may not move the needle in terms of brand recall.

Awareness tactic: optimizing for website engagement over traffic.

It takes a lot more involvement from a user to load your website and then scroll down it or click on certain areas firing events than to just click on the ad link.

If you’ve tried this before and your CPC was significantly more expensive optimizing for website engagement than for your traffic campaign, you can test more creative assets to try lowering CPC or test optimizing for a website event that will provide a higher number of signals to pass back to the ad platform acquisition algorithm.

Also, I’d recommend setting up the campaigns in a way that enables you to track the average bounce rate of each campaign; website engagement vs. pure traffic. Typically, I find that the pure traffic campaigns have a higher bounce rate and that actually offsets any increases in average CPC which in theory makes the website engagement campaign actually provide cheaper traffic.

*Note, this can work for any popular ad platform including Facebook/Instagram, Tik Tok, Google, Pinterest, etc…

But remember, an algorithm is only as good as its data.


By creating more data points to measure with improved awareness (e.g. Top of Funnel) campaigns, you can better forecast the customer lifetime value by analyzing the traits of your highest LTV customers and focusing your consideration campaigns on them.

Can you find patterns that all of your highest LTV customers have in common? Perhaps they all purchase similar products or view similar content on your website.

For your consideration campaigns, you can go one step further and create content which speaks specifically to the products that your highest LTV customers purchase and then work to find the right messaging to encourage prospective users to consume that content and further their journey to becoming another one of your top LTV customers.

In addition, if you have products or services that are very high LTV and very low LTV, you can work to sell the low LTV items as lead items which produce a low or breakeven margin. You may find that by moving these items into your “consideration” or MOF budget, you’re able to run your ads at a higher ROI when comparing against your total revenue because you’ve enabled your email, SMS or 1-on-1 sales tactics to gain higher LTV sales.

If you take a step back from your funnel, the goal of your top and middle areas should be to gain enough data points that you can really understand your audience. Once you understand them, you can provide more value to them.


In this segment, we should work to find causal relationships between the data signals we’ve gained in the top or middle of the funnel and what ends up providing the highest LTV customers for your brand.

With a well thought out A/B testing matrix, we can test to see if our hypotheses are true.

Here are some potential causal relationships you may want to look into for your brand:

  • Reassurance messaging. What messaging is most likely to create trust with your highest LTV customers. Is it the fact that you have hundreds of positive reviews, or that you provide a lifetime warranty, or that you have exceptional post-purchase customer success? Test this, find a causal relationship so you know what messaging you should have in the most prominent places along your journey.

  • Previous products purchased. What products have driven repeat purchases and have the lowest rate of returns? Find out so you can double down on improving your marketing materials and competitive advantage on those products.

  • Previous content consumed. What videos, ads, articles, emails or support inquiries do your highest LTV customers have in common? If you’re able to pinpoint this, you’ll be able to ensure more prospective customers get the same priming.

In sum, I hope this article has inspired you to think more critically about how you’re running your paid media campaigns in each layer of the funnel. If you’re looking for any support, don’t hesitate to reach out. If you can’t tell, I’m very passionate about all things performance marketing related and always love to pass on any advice I can.

Happy testing,



Build a winning growth strategy.

Free digital marketing insights by our CEO.

Never miss an update

Thanks for submitting!

bottom of page