You might not know it, but your entire digital ad program is likely one big remarketing campaign.

I know what you’re thinking: "How could that be? We aren't running specific remarketing campaigns on any of our platforms."

The culprit isn’t your strategy, but the reality of how modern bidding algorithms work, and why they struggle to deliver consistent growth.

The algorithm’s dilemma

Tools like Meta Bid Strategies and Google Smart Bidding have a single mandate: hit the tCPA (Target Cost Per Acquisition) or tROAS (Target Return on Ad Spend) targets you set. They have to figure out which impression will generate an ad conversion, and they have to do it for millions of advertisers simultaneously.

We assume platforms use countless signals to acquire customers for each unique advertiser — in reality, they can’t run millions of perfectly unique bidding algorithms. They need features and attributes that generalize, and there’s one data point that generalizes better than any other: Has this person been to the advertiser's website before?

The path of least resistance

No matter what new keywords you test, what creatives you launch, or what audiences you add, the automated bidder eventually faces a binary choice between two otherwise identical impressions:

Show an impression to the person who has never heard of you.
Show an impression to the person who has visited your site before.

If the goal is to observe a conversion as fast as possible, the algorithm picks the past visitor nearly every single time.

The "scale" wall

Have you ever launched a campaign where you initially see a big business impact, but eventually, it struggles to scale and eventually taps out?

This explains a phenomenon almost every marketer has faced, and occurs whenever algorithmic bidding trains on ad platform conversion data.

Why? Because over millions of those binary decisions, the algorithm locks into hunting only past visitors.  Effectively, given enough time and money, every well-crafted prospecting or acquisition campaign turns into a glorified remarketing campaign.

How to fix it

So, what do you do? You can’t stop running digital marketing — 80% of sales involve a digital touchpoint before the purchase. Giving that up means giving up the majority of your revenue growth opportunity.

The solution lies in automating your bidding strategy without relying on ad platform conversion data.

If you feed the algorithm the standard conversion signals, it will gradually enhance remarketing behavior. By changing the data you bid against, you can force the platform to reach the net-new audiences you actually want.

Replace conversion data with predicted click-value. Plug it directly into Google Smart Bidding and Meta Bid Strategies. By doing so, you’ll break the remarketing doom-loop, and reach the audiences you’re actually targeting.

This approach is why Bonsai sees case studies with ridiculous numbers, such as 50% business growth. I’m not talking about "platform conversions" that look good on a dashboard, either.  I’m talking about a company’s total top-line revenue.

Just like click-through rate prediction (pCTR) made CPC ads possible, click-value prediction models make CPC ads profitable and scalable once again.