Taboola Audiences for Native Ads: The Broad-First Play (2026)

Most advertisers switch to Taboola from Facebook or TikTok, go hunting for interest targeting on day one, and quietly bleed margin for weeks. The audiences exist on Taboola. The mistake is bolting them onto a brand-new account before you have a single dollar of conversion data to back the choice.

This is the exact play I use to turn Taboola audiences into a profit lever instead of a leak: start broad, spend into real data on accounts running $50-100K, then carve out the segments hitting a $13 CPA and build dedicated campaigns around them.

Marcel Sattler is the founder of native-advertising.net, and since 2015 he has deployed more than $100M across Taboola, Outbrain, Newsbreak, MGID, Yahoo Native, Mediago, and RevContent in DTC, lead-gen, and affiliate. Audience targeting is one of the most misused tools on Taboola, so here is how I actually run it.

Why Taboola audiences are not Facebook interests

If you come from social, you expect to type "interested in honey" or "back pain products" and get a tight, precise segment. That is not what Taboola gives you. Taboola audiences are built around the kind of pages people read — interested in health, interested in motorcycles — not granular declared interests.

That distinction matters for how hard you lean on them. A Facebook segment like "relationship status: single" is precise. The equivalent on Taboola is not. So treat these audiences as directional signals, not surgical filters.

There are two flavors to work with: Taboola's own first-party audiences and external audiences, many of them from data providers like Oracle. You will see names like "Shopping Holiday and Special Event Shoppers by Oracle" or "retail purchase-based Home and Garden buyers." Useful, but fuzzy.

One more reality: audiences overlap. The same user can sit inside three or four segments at once. That overlap is exactly why evaluating them on a clean broad baseline beats guessing up front.

Why you start broad, not with audiences

Native advertising is a scaling game. When you run with no targeting, you have close to unlimited scaling room — Taboola can keep feeding you reach as long as the campaign performs. The moment you bolt on an audience, you cap that reach.

So the funnel starts wide on purpose. Launch broad: no audience targeting, no interest layer, nothing. Let the platform spend and let conversions accumulate against the full inventory. That broad campaign becomes the data engine for everything you build later.

If you have read my other breakdowns, you know I push broad-first on native for a reason. Targeting too early starves the algorithm and hands you a small, biased sample. For DTC and dropship accounts, that early cap on scale is the difference between a winner and a stalled test — which is why the ecommerce play almost always opens broad.

How much you need to spend before audiences mean anything

This is where most people get it wrong. They drop $5K or $10K into an account, open the audience report, and decide they have "all the insights." They do not.

For audience data to be trustworthy, you are looking at accounts spending real money — at least $50K, $80K, $100K per account before the audience-level numbers stabilize. Below that, you are reading noise.

There is a per-audience threshold too. Inside the report, you want significant spend concentrated on each audience you intend to act on — roughly $8K to $10K minimum on a single audience. I have seen audiences with $25K, $20K, $8K, and $10K of spend each in a client account, and that is the kind of volume that makes a CPA number mean something. Three hundred dollars on an audience tells you nothing.

This is why the play lives on funded accounts. A lead-gen or affiliate account scaling past $50K is where audience analysis earns its keep.

Reading the Taboola audience report

Log into your Taboola account and find the audience section in the top menu. By default it filters to the top 200 audiences. Scroll and you will see the partner column — whether it is Taboola itself or a data provider like Oracle — plus the audience name, spend, impressions, and the number that actually decides everything: CPA.

Your job in this report is simple. Find the cheapest CPAs.

• Partner — tells you if the segment is Taboola's first-party data or an external provider.
• Audience name — read it for intent ("Home Improvement," "Insurance," "Home and Garden buyers").
• Spend — only trust audiences with $8-10K behind them.
• CPA — the deciding metric; rank from cheapest to most expensive.

Remove the default filter so you can see the full weight distribution. The range is wide. In one client account, the best audiences were converting at a $13 CPA while others sat far higher. You are not trying to use them all — you are isolating the cheap, high-relevance segments and ignoring the expensive tail.

Building the audience campaigns

Once the broad campaign has spent into real data and you have your shortlist of cheap-CPA, high-spend audiences, you build from scratch. Do not retrofit audiences onto the broad campaign. Create new, dedicated campaigns that contain only the audiences converting best.

1. Run broad with no targeting and let it scale.
2. Spend into significance — $50-100K at the account level, $8-10K per audience.
3. Pull the audience report and rank by CPA, cheapest first.
4. Confirm each finalist has enough spend behind it to be real.
5. Build new campaigns from scratch around only the winning audiences.

These audience campaigns will be smaller. That is expected and fine. You trade reach for a higher conversion rate. The broad campaign keeps the scale; the audience campaigns squeeze out efficiency on the segments you have proven.

This pattern travels across networks. The same broad-first-then-segment logic is how I approach optimization whether the spend is on Taboola or Outbrain.

Do not hyper-focus on audiences

Here is the recommendation I hammer on every time: audiences are one optimization lever, not the whole game. Build a combination. Run broad campaigns, audience campaigns, and retargeting campaigns side by side. Target different device types. Slice by browser, by GEO, by region.

Audience optimization is one type of optimization among many. If you tunnel-vision on audiences and ignore the rest, you leave just as much performance on the table as the people who launched audiences too early.

The strongest accounts I run keep all of these levers in motion at once — and the audience work only matters because it sits on top of clean broad data and a structure that is already scaling. Audiences in the US, for what it is worth, consistently outperform every other country I have tested them in, so weight your effort accordingly.

Watch the full breakdown

Is your account a fit for the same play?

If your account is already spending $50K or more a month on native and you suspect your audiences are capping scale instead of lifting profit, this is a structure problem worth a second set of eyes. The broad-first funnel only pays off when the data underneath it is clean and the spend is concentrated enough to trust.

Book a strategy call and we will look at your current Taboola structure together. If you want proof first, the case studies show the outcomes, and the full resource library has the rest of the native playbook — including the affiliate-specific version of this for affiliate accounts.