Listen to this article · 12 min listen

Key Takeaways

  • Advertisers who effectively use custom audiences for video ads see a 3x higher return on ad spend (ROAS) compared to those relying solely on broad targeting.
  • Implementing lookalike audiences based on a seed list of high-value customers can expand reach by an average of 15% while maintaining conversion efficiency.
  • Retargeting site visitors who watched 75% or more of a previous video ad yields a 20% higher click-through rate (CTR) than general site visitor retargeting.
  • Excluding irrelevant audience segments, such as recent purchasers for acquisition campaigns, can reduce wasted ad spend by up to 30%.
  • A/B testing at least three distinct custom audience segments per video ad campaign provides actionable insights for optimization within the first two weeks of launch.

Did you know that 68% of consumers prefer watching short videos to learn about a new product or service, yet only a fraction of businesses truly master custom audiences for video ads? That’s a staggering missed opportunity. Precision targeting isn’t just a nice-to-have; it’s the bedrock of any successful video advertising strategy in 2026, transforming casual viewers into loyal customers.

Only 15% of Marketers Consistently Segment Video Ad Audiences Beyond Basic Demographics

This statistic, from a recent eMarketer report, is frankly astonishing. It tells me that most advertisers are leaving immense value on the table. When I first started in this field, we were happy just getting video in front of eyeballs. Now, with the sophisticated tools available, relying on age, gender, and location alone for video ads is like trying to catch fish with a sieve. You’ll get some, sure, but you’ll miss most of the good ones. My own experience bears this out: I had a client last year, a niche e-commerce brand selling artisanal coffee equipment, who was running broad demographic campaigns on a popular video platform. Their cost per acquisition (CPA) was through the roof. We implemented custom audiences based on website visitor behavior, specifically targeting users who had viewed product pages for brewing equipment but hadn’t purchased. Within three weeks, their CPA dropped by 40%. It wasn’t magic; it was just common sense applied with precision.

The conventional wisdom often pushes for “maximum reach” with video, assuming that if enough people see it, some will convert. I disagree fundamentally. For video ads, especially, quality of reach trumps quantity every single time. A highly engaged, smaller audience who watches your video because it’s directly relevant to their interests will always outperform a massive, indifferent audience. Think about it: a person actively researching espresso machines is far more likely to respond to a video ad about a new grinder than someone who just happens to be in the right age bracket. This isn’t just about efficiency; it’s about building a better brand experience. Nobody likes irrelevant ads, and video ads, by their nature, demand more attention. Wasting that attention on the wrong people is a cardinal sin.

Video Ads Targeting Custom Audiences Show a 2.5x Higher Engagement Rate Than Broad Campaigns

This figure, often cited in Google Ads documentation, highlights the power of relevance. Higher engagement means more completed views, more clicks, and ultimately, a stronger connection with your potential customer. When we talk about engagement in video, we’re not just looking at views; we’re analyzing view-through rates, click-through rates (CTR), and even post-view conversions. If your video ad is showing up in front of someone who has already expressed interest in your product category, they’re far more likely to watch it to completion. They’re also more likely to click through to your landing page. This is where the magic of audience segmentation truly shines.

We’ve found that creating custom audiences based on various signals is key. This includes:

  • Website Visitors: Segmenting by specific pages visited, time spent on site, or even actions taken (e.g., added to cart, viewed a demo).
  • Customer Lists: Uploading email lists of existing customers or leads for retargeting or exclusion.
  • App Users: For mobile-first businesses, targeting users based on app installs, in-app purchases, or specific feature usage.
  • Engagement Audiences: People who have previously interacted with your social media content or watched other videos from your brand.

My team recently worked on a campaign for a B2B SaaS company launching a new feature. Instead of targeting all their existing customers, which would have been a waste for those not using the relevant part of the software, we created a custom audience of current users who had interacted with the old version of that feature. The video ad showcasing the new functionality had an incredible 70% view-through rate and a CTR three times higher than their previous broad-reach product announcement videos. It just goes to show: speak directly to the choir, and they’ll sing your praises.

Lookalike Audiences Generated from Top 10% Converters Outperform Standard Lookalikes by 30%

The conventional wisdom around lookalike audiences is often “just use your customer list.” While that’s a good start, it’s not enough. We’ve consistently seen that refining your source audience for lookalikes dramatically improves performance. A recent IAB report highlighted the increasing sophistication of lookalike modeling, and this statistic resonates deeply with my practical experience. Creating lookalikes from your entire customer base includes low-value customers, one-time purchasers, or even those who converted but then churned quickly. These aren’t the people you want to clone.

Instead, I advocate for building your lookalike seed audiences from your most valuable segments:

  • Your top 10% of customers by lifetime value (LTV).
  • Customers who have made repeat purchases.
  • Users who have completed a specific high-value action, like a demo request or a free trial signup.
  • Website visitors who spent an unusually long time on key conversion pages.

This approach ensures that the algorithm has the richest possible data points to find new users who are genuinely similar to your best existing ones. We ran an experiment for a financial services client where we compared two lookalike audiences for a video ad promoting a new investment product. One was built from their entire client list, the other from clients who had invested over a certain threshold and maintained their account for more than three years. The “high-value” lookalike audience had a 30% lower cost per lead (CPL) and a significantly higher lead-to-client conversion rate. This wasn’t a minor tweak; it was a fundamental shift in strategy that paid dividends.

Here’s an editorial aside: many marketers get lazy with lookalikes. They set them up once and forget them. That’s a mistake. Your best customers evolve, your market shifts. You need to refresh your lookalike seed audiences regularly, at least quarterly, to ensure they remain potent. Don’t just set it and forget it; analyze it and refine it.

Excluding Past Converters from Acquisition Campaigns Reduces Wasted Ad Spend by an Average of 25%

This might seem incredibly obvious, but you’d be shocked how often I see businesses pouring money into showing acquisition-focused video ads to people who have already bought their product. It’s an easy win, yet it’s frequently overlooked. A HubSpot report on marketing efficiency emphasized the importance of negative targeting, and for good reason. Why pay to acquire a customer you already have? It’s not just wasted spend; it’s also a poor customer experience. Imagine seeing an ad for a product you just bought last week. Annoying, right? Video ads, being inherently more intrusive, amplify this annoyance.

Effective exclusion lists are just as important as your inclusion lists. For video ads, this means:

  • Recent Purchasers: Exclude anyone who has converted in the last 30, 60, or 90 days (depending on your sales cycle).
  • Existing Customers: If the video ad is for a new customer offer, exclude your entire customer database.
  • Unqualified Leads: If you have a CRM with lead statuses, exclude those marked as “disqualified” or “unresponsive.”
  • Specific Content Viewers: For retargeting, exclude those who have already watched 100% of a specific promotional video if your goal is to show them something new.

We ran into this exact issue at my previous firm with a subscription box service. Their acquisition video ads were showing up repeatedly for existing subscribers. We implemented an exclusion audience based on their subscriber list, and within the first month, they saw a 28% reduction in their overall video ad spend for acquisition campaigns, with no dip in new sign-ups. That’s money directly back into their pocket, or better yet, reallocated to more effective strategies. This isn’t just about saving money; it’s about respecting your audience’s time and attention. Showing them relevant content, or avoiding irrelevant content, builds goodwill.

The Average Advertiser Uses Only Two Custom Audience Types for Video Ads, Missing Out on Deeper Personalization

This observation isn’t from a single report, but rather a pattern I’ve seen across hundreds of ad accounts I’ve audited over the years. Most marketers stick to the basics: website visitors and perhaps a customer list lookalike. While these are foundational, they barely scratch the surface of what’s possible with video ad targeting. The platforms themselves, like Meta Business Help Center, offer a wealth of custom audience options that go largely underutilized.

Consider the potential of combining audience types:

  • Video Viewers + Website Visitors: Target people who watched 75% of your previous video ad AND visited a specific product page. This signals extremely high intent.
  • Customer List + Engagement Audience: Upload a list of inactive customers and target them with a video ad showing a new product feature, but only if they’ve also engaged with one of your social posts recently. This prioritizes those who are still somewhat connected.
  • App Users + Geo-targeting: For a retail brand, target users of their shopping app who are currently within a 5-mile radius of a new store opening with a video ad promoting in-store exclusive deals.

I had a client who sold high-end fitness equipment. They were running a general retargeting campaign for all website visitors. I argued that we needed to get more granular. We created a custom audience of users who had viewed videos on their site about specific equipment (treadmills, ellipticals, weights) AND then visited the corresponding product pages but hadn’t purchased. We then served them highly specific video ads featuring those exact products. The result? A 5x increase in conversion rate for that segment. It proved to me that specificity isn’t just good; it’s transformative. This level of granular targeting allows you to craft video ad creative that speaks directly to a person’s immediate needs and interests, making your message far more impactful.

In 2026, relying on broad strokes for your video advertising is a recipe for mediocrity. By embracing sophisticated custom audiences and continuously refining your audience segmentation, you can transform your video ad performance from merely acceptable to truly exceptional, delivering precise messages to the right people every single time.

What is a custom audience in video advertising?

A custom audience in video advertising is a targeted group of people identified by specific data points, such as website visits, customer lists, app usage, or engagement with your content, allowing advertisers to serve highly relevant video ads to them. This moves beyond basic demographics to focus on demonstrated interest or existing relationships.

How do I create an effective custom audience for my video ads?

To create an effective custom audience, start by defining your campaign objective (e.g., brand awareness, lead generation, sales). Then, gather data from your customer relationship management (CRM) system, website analytics, or app usage. Segment this data based on behaviors like product page views, video watch times (e.g., 75% completion), or past purchase history. For lookalikes, use your highest-value customer segments as the seed audience.

What is the difference between custom audiences and lookalike audiences?

Custom audiences are built from your existing data (e.g., website visitors, email lists). Lookalike audiences are created by advertising platforms using your custom audience as a “seed” to find new users who share similar characteristics and behaviors to your existing valuable customers or leads. Lookalikes expand your reach to new, relevant prospects.

Can I use custom audiences to exclude certain people from seeing my video ads?

Yes, absolutely. Exclusion targeting is a critical component of custom audiences. You can create custom audiences of people you do not want to reach with a specific video ad campaign, such as recent purchasers for an acquisition campaign, or existing customers for a new customer offer. This prevents wasted ad spend and improves the customer experience.

How often should I update my custom audiences for video ad campaigns?

The frequency of updating custom audiences depends on the dynamism of your business and customer lifecycle. For highly active websites or fast-moving product cycles, daily or weekly updates are ideal for website visitor segments. Customer list-based audiences or lookalikes should be refreshed at least quarterly, or whenever there’s significant change in your customer base, to ensure accuracy and optimal performance.