If your retargeting strategy still looks like it did three years ago, you are almost certainly leaving money on the table and paying more than you should to acquire every customer.
The old playbook was simple enough: install the pixel, build a website visitor audience, show them the same product they looked at, repeat until they buy or opt out. For a while, it worked well. Platforms had rich data, tracking was reliable, and the audiences were big enough to be meaningful.
Then privacy changes came. iOS updates gutted pixel data. Cookies started disappearing. Audience sizes shrank. And the brands still running the old playbook watched their retargeting performance quietly deteriorate while their customer acquisition costs crept upward.
The good news is that retargeting has not stopped working. It has evolved. The brands seeing the strongest return on their ad spend right now are the ones that have adapted their mid-funnel strategy to match how platforms actually function today, and how buyers actually make decisions. This post breaks down what that looks like and what you can start implementing now.
Why the Old Retargeting Playbook Broke Down
The core problem is data loss. Apple's App Tracking Transparency framework, introduced in 2021, gave iPhone users the ability to opt out of cross-app tracking. The majority of them did. That single change eliminated a significant portion of the behavioral data that platforms like Meta had been using to build retargeting audiences and optimize ad delivery.
At the same time, browser-level cookie restrictions tightened across Chrome, Safari, and Firefox, further reducing the signal available to advertisers who relied on third-party tracking.
The result for businesses running standard pixel-based retargeting was smaller audiences, less accurate attribution, and diminishing performance at the same or higher spend levels. Cost per acquisition rose not because the audiences were less interested, but because the platforms had less information to work with.
The response cannot be to simply spend more on a broken model. It has to be to rebuild the mid-funnel strategy on a foundation that works under current conditions.
What the Mid-Funnel Actually Means Now
The mid-funnel used to be a fairly narrow band: people who had visited your website and not yet purchased. That was the retargeting pool, and it was defined almost entirely by pixel data.
Today, the mid-funnel is better understood as anyone who has had a meaningful interaction with your brand, regardless of whether that interaction happened on your website. That definition opens up a much wider set of audiences and a more flexible set of strategies.
Warm audiences in 2025 include people who have:
- Watched a significant portion of your video content on Meta, TikTok, or YouTube
- Engaged with your organic social posts, reels, or stories
- Clicked on a previous ad without converting
- Opened or clicked your emails
- Submitted a lead form or started a checkout
- Been uploaded as a customer or prospect list from your CRM
- Been matched through server-side or first-party data signals
When you think about the mid-funnel this way, your warm audience is substantially larger than your pixel-based website visitor pool. And because these audiences are built on first-party and platform-native signals rather than third-party cookies, they are also more durable as privacy standards continue to tighten.
The Strategies That Are Actually Working Right Now
1. Video View Audiences as a Retargeting Foundation
Video engagement is one of the most powerful and underused mid-funnel signals available. On Meta and TikTok, you can build audiences of users who watched 25%, 50%, 75%, or 95% of any video in your ad account. These are people who chose to spend time with your content. That is a meaningful signal of interest.
The practical application: run broad, awareness-level video content at the top of funnel to generate video views, then retarget that engaged audience with a more direct conversion message. The sequence feels natural to the viewer and gives the platform enough engagement data to optimize effectively, even with reduced pixel signal.
Brands running this structure consistently see lower cost per acquisition compared to pixel-only retargeting, because the audience quality is high and the platform has clear engagement signals to work with.
2. Meta Engagement Custom Audiences
Beyond video, Meta lets you build retargeting audiences from users who have interacted with your Facebook page, Instagram profile, or individual posts and ads. This includes people who liked, commented, shared, saved, or sent a message in response to any of your content.
These audiences are built entirely on platform-native data, which means they are unaffected by iOS tracking restrictions. And because they represent people who have already demonstrated some level of interest in your brand, they tend to convert at a significantly higher rate than cold audiences.
If you are not actively building and testing Meta engagement audiences as a core retargeting layer, you are missing one of the most accessible warm audience strategies available right now.
3. First-Party Data and CRM Uploads
Your customer and prospect list is one of the most valuable assets in your marketing operation, and most businesses are not fully leveraging it for paid media.
Uploading your email list to Meta, Google, and TikTok allows you to target existing customers with retention and upsell campaigns, suppress recent buyers from prospecting campaigns to avoid waste, and build lookalike audiences modeled on your best customers.
The accuracy of matched audiences depends on the quality and completeness of your list, which is one reason why collecting first-party data, including email, phone number, and name, should be a consistent priority across every customer touchpoint. Businesses that have been building and maintaining clean CRM data are at a significant advantage right now compared to those that have relied primarily on pixel tracking.
4. Server-Side Tracking and the Conversions API
One of the most impactful technical improvements a business can make to recover lost signal is implementing server-side tracking alongside or in place of standard browser-based pixel tracking.
Meta's Conversions API (CAPI), Google's enhanced conversions, and similar tools from TikTok allow conversion events to be sent directly from your server to the ad platform, bypassing browser-level restrictions entirely. This recovers a meaningful portion of the data that iOS and browser changes eliminated.
Businesses that have implemented server-side tracking typically see improved attribution, better optimization, and more accurate audience building across all their campaigns, not just retargeting. It is one of the highest-return technical investments available in paid media right now.
5. Sequential Messaging Across the Funnel
One of the consistent mistakes in retargeting is showing the same message to everyone in the warm audience regardless of where they are in the decision process. Someone who watched a 15-second video clip and someone who added a product to cart and abandoned it are at very different points in their buying journey. They should not be seeing the same ad.
Sequential messaging means designing a deliberate series of ad experiences that move a prospect forward based on what they have already seen and done. A typical sequence might look like this:
- Stage 1 (first exposure): Brand story, product introduction, or educational content that builds familiarity
- Stage 2 (after engagement): Social proof, testimonials, or use-case specific content that addresses hesitation
- Stage 3 (high intent signals): Direct offer, limited-time incentive, or urgency-based creative for users who have taken a high-intent action like visiting a product page or initiating checkout
This approach respects where the buyer actually is, reduces ad fatigue, and consistently outperforms single-message retargeting campaigns in both conversion rate and cost per acquisition.
6. Lookalike Audiences Built From Warm Signals
Lookalike audiences have always been part of the paid media toolkit, but the seed audiences you use to generate them make a significant difference in performance. The best lookalikes today are built not just from customer purchase lists, but from high-quality engagement signals.
Strong seed audiences for lookalike generation include your top 10% to 25% of customers ranked by lifetime value, users who completed a high-intent action like a consultation booking or checkout initiation, video viewers who watched 75% or more of your content, and email subscribers with a strong open and click history.
When the seed audience is well-defined and genuinely represents your best buyers, the lookalike audiences generated from it tend to be substantially stronger than those built from broad pixel-based pools.
How This Connects to Lower Customer Acquisition Costs
Every one of these strategies is ultimately about the same thing: getting your message in front of people who already have some familiarity with your brand, using better-quality signals to find them, and delivering messaging that matches where they are in the buying process.
When you do those three things well, conversion rates go up and the cost to acquire each customer goes down. You are not spending money to convince cold strangers. You are spending money to follow up with people who have already raised their hand in some way. That is inherently more efficient.
The businesses consistently achieving the lowest customer acquisition costs in paid media are not necessarily the ones with the biggest budgets. They are the ones with the most sophisticated audience strategy, the cleanest first-party data, and the most intentional mid-funnel creative sequences. Budget amplifies strategy. It does not replace it.
What to Audit in Your Current Retargeting Setup
If you want to assess where your current mid-funnel strategy stands, start by working through these questions:
- Are you building and testing video view audiences, or relying primarily on website visitor pixel data?
- Are you using Meta and TikTok engagement audiences as a retargeting layer?
- Is your CRM list being uploaded and regularly refreshed in each ad platform?
- Have you implemented server-side tracking through the Meta Conversions API, Google enhanced conversions, or equivalent?
- Are you showing the same creative to all warm audiences, or sequencing messages based on engagement level?
- Are your lookalike audiences being seeded with your highest-quality buyers, or just general customer lists?
If the answer to most of those questions is no, or you are not sure, there is meaningful room to improve performance without increasing spend. In many cases, restructuring the mid-funnel strategy alone is enough to drive a measurable reduction in customer acquisition cost within the first 60 to 90 days.
Ready to Rebuild Your Mid-Funnel Strategy?
Retargeting that worked three years ago is not the same as retargeting that works today. The businesses growing efficiently on paid media right now have adapted their approach to match the current data environment, and they are seeing lower CAC and stronger ROAS as a result.
At Soda Spoon Marketing, we help local service businesses, B2B companies, and e-commerce brands build paid media strategies that actually perform in today's environment, from first-party data infrastructure to full mid-funnel creative sequences across Meta, Google, and TikTok.
