How to find and block the Taboola publisher sites that waste your budget

12 min readUpdated Nov 8, 2026

How to find and block the Taboola publisher sites that waste your budget

If you want to Taboola block sites effectively, stop guessing from the Taboola dashboard alone. The sites that waste your budget are the ones that send clicks but never send conversions — and the fastest way to find them is to pass Taboola's site token into a sub-ID, record conversions by postback, and read your own per-site data. Then you block in Taboola based on evidence, not instinct. This guide walks through the exact setup, the thresholds to use, and the loop that keeps your site list clean as campaigns scale.

Everything below assumes you're running native campaigns on Taboola and tracking them in a tracker you control. If you haven't wired the network up yet, start with How to track native and push ad campaigns: a network-by-network setup guide and come back here once clicks and conversions are landing in the same place.

Why does Taboola's own reporting not tell you which sites to block?

Taboola reports on campaigns, ad items and — depending on your account setup — publisher breakdowns. What it can't do is reconcile spend against the conversions your affiliate link actually produced. That reconciliation is the whole game, because a publisher site is only "bad" if it spends money without producing revenue.

There are three specific gaps:

  1. Click counts differ. Taboola counts a click when its redirect fires. Your tracker counts a click when the landing page request arrives. Those are different events and they will never match exactly. This is normal and covered in Why your ad network and your tracker show different click counts.
  2. Conversions live downstream. If your offer converts on a third-party network or your own checkout, Taboola has no visibility into it unless you send it back. That requires a postback, covered in Taboola conversion tracking for affiliates: sending sales back with a postback.
  3. Per-site granularity is limited. You need per-site spend, per-site clicks and per-site conversions in one table. That's a tracker job, not a network dashboard job.

The practical consequence: the site you should block is the one where your tracker shows spend with zero or near-zero conversions after a statistically meaningful number of clicks. You can only see that if Taboola's site identifier is flowing into your tracker on every click.

Step 1: Pass the Taboola site token into a sub-ID

Taboola exposes a set of tracking macros you can place in your campaign URL. The one that matters here is the publisher/site token — the macro that resolves to the site your ad appeared on. The exact macro name is documented in Taboola's own tracking macro reference; check it before you build, because networks rename and add tokens over time.

Your campaign URL should look structurally like this:

https://yourdomain.com/click?campaign=tab_native_us_01&sub1={site_macro}&sub2={campaign_id_macro}&sub3={ad_id_macro}

Where {site_macro} is Taboola's publisher site token, {campaign_id_macro} is the campaign identifier and {ad_id_macro} is the creative/ad item identifier. Map them consistently:

  • sub1 → publisher site (the thing you're blocking)
  • sub2 → Taboola campaign ID
  • sub3 → ad item / creative ID
  • sub4, sub5 → reserve for landing page variant and traffic segment so you don't have to rebuild URLs later

If you're unsure which token is which, Ad network tracking macros explained: what tokens are and how to map them to sub-IDs covers the general model. The rule is simple: one token per sub-ID, documented in a spreadsheet, never reused for two meanings.

On Adtraxo, sub1–sub5 are Pro-plan features, so this setup requires Pro. The free plan gives you 10 tracking links, 5,000 recorded clicks a month and basic stats, but not sub-IDs or postback conversions — you need Pro ($27/month at the founding rate) for the per-site view this article depends on.

Step 2: Send conversions back by postback

A site list built on clicks is useless. You need conversions attributed to the same sub1 value that carried the site token. That means a server-to-server postback from your offer source or affiliate network into your tracker, with the click ID passed through.

  1. In your tracker, create the conversion event and copy the postback URL it generates.
  2. Paste that URL into your affiliate network or offer platform's postback settings, with their click-ID macro in the placeholder.
  3. Fire a test conversion and confirm it appears against the correct click — and therefore the correct sub1 site.
  4. Confirm the conversion shows the same sub1 value as the click record. If it doesn't, your click ID isn't being passed end to end.

Don't skip the test. A postback that fires but loses the click ID will attribute conversions to nothing, and your per-site report will show every site as a zero-conversion site — which is worse than no report at all.

Step 3: Read the per-site report and decide what to cut

Now open your tracker's per-campaign breakdown and group by sub1. You're looking at four columns per site: clicks, cost, conversions, and revenue (or payout). Everything else is commentary.

Classify each site into one of four buckets:

  1. Converters. Conversions at or above your target CPA or ROI. Keep, and consider scaling spend.
  2. Undecided. Some clicks, no conversions yet, but not enough volume to conclude anything. Leave running.
  3. Bleeders. Spend above your per-site tolerance with zero conversions after enough clicks to be meaningful. Block.
  4. Zombies. Tiny trickle of clicks, tiny spend, no conversions. Block on housekeeping grounds — they add noise to every report.

The "enough clicks" threshold is the part everyone gets wrong. Don't pick a round number because it feels safe. Compute it from your own economics:

Minimum clicks before judging a site ≈ (cost you're willing to lose testing one site) ÷ (your average CPC on that campaign)

Worked example: your Taboola campaign averages $0.30 CPC and you're willing to lose $15 testing any single site. That's 50 clicks. A site with 50 clicks and zero conversions gets blocked. A site with 12 clicks and zero conversions stays in "undecided" — you haven't spent enough to learn anything.

If your offer converts at roughly 1 in 100 clicks on good traffic, a site with 50 clicks and no conversions is weak evidence, not proof. That's why the tolerance figure should reflect what you can afford to lose, not what feels conclusive. Raise the tolerance when you have budget to learn; lower it when you're protecting a thin margin.

Step 4: Block the site in Taboola, not just in your tracker

Blocking inside your tracker only stops the reporting from being polluted. The money keeps flowing until you block at the source. In Taboola's campaign settings there is a site/publisher exclusion list; add the domain there.

  1. Export the list of blocked sites from your tracker (or copy the sub1 values).
  2. Open the campaign's site exclusion settings in Taboola.
  3. Paste the domains in, one per line, and save.
  4. Note the date and the reason in a change log — you'll want to revisit blocks in 60–90 days.
  5. Apply the same list to sibling campaigns targeting the same geo and vertical, since bad sites tend to repeat.

Two practical notes. First, block at the campaign level for speed, then consolidate into an account-level exclusion list once the pattern is clear. Second, keep a "review later" tab. Sites that were blocked for zero conversions in a cold season sometimes perform in a warm one — but only revisit with fresh data, not sentiment.

Step 5: Close the loop so the list stays clean

A one-off block list rots. New sites enter Taboola's inventory constantly, and a campaign that was clean last month can pick up bleeders this month. Build a weekly rhythm:

  1. Monday: pull the per-site report for every active Taboola campaign, sorted by cost descending.
  2. Flag: mark every site above your tolerance with zero conversions.
  3. Block: add flagged domains to the campaign exclusion list and the shared block list.
  4. Promote: identify sites with strong conversion rates and either raise bids or move them into a dedicated campaign so you can control their budget separately.
  5. Log: record what you blocked, why, and the date. This is the only way to audit your own decisions later.

If you want the same discipline applied to other networks in your stack, the pattern is identical: MGID widget blacklist: how to decide which widgets to cut and when and How to build a PropellerAds zone blacklist from your own tracker data walk through the equivalents. The mechanics differ; the decision rule doesn't.

What else should you check before you block a site?

Before you cut, rule out the four things that make a good site look bad:

  • Broken postback. If conversions stopped appearing for all sites at a specific timestamp, the problem is your tracking, not the traffic. Check the postback log first.
  • Landing page mismatch. A site can be fine but your LP copy can be wrong for its audience. If you're split-testing landing pages, check whether the zero-conversion site was only ever served the losing variant.
  • Creative fatigue. A site with a single stale creative may look dead. Test a fresh creative before blocking — see Push ad creative testing: how to compare creatives by conversions, not clicks for the method.
  • Fraud signals. If a site shows high clicks, near-zero landing page arrivals, and suspicious IP patterns, it may be bot traffic rather than merely poor traffic. Adtraxo's fraud detection (IP velocity, datacenter IP, bot user agents, uniqueness conflicts, invalid referers) surfaces this on Pro, and those sites should be blocked immediately, not tested further.

Also compare click counts between Taboola and your tracker per site. A site where Taboola reports 500 clicks and your tracker records 40 landing page arrivals is a red flag — either the redirect is failing or the traffic isn't human. The arrival snippet on your landing page (available on any Adtraxo plan) tells you which clicks actually loaded the page in a real browser, which is the cleanest way to separate "bad traffic" from "bad tracking".

How many sites should you expect to block?

There's no universal number, and anyone quoting one is guessing. What you can control is the process: a per-site view, a tolerance threshold derived from your own CPC and risk appetite, and a weekly review. In practice, most native campaigns end up with a block list that grows quickly in the first two weeks and then settles into a slow trickle as the obvious waste is removed. If your block list never grows, you're probably not looking at the data closely enough — or your tolerance is set so high you're only catching catastrophic sites.

The other mistake is over-blocking. Cutting a site after 15 clicks because it "feels bad" removes the traffic that would have converted at scale. Set the threshold before you look at the report, and stick to it.

Can you automate the decision?

You can automate the flagging, and you should. The manual part is the judgement call on tolerance and on whether a zero-conversion site is a tracking failure or a genuine bleeder. Adtraxo's AI Campaign Optimizer agent works on top of the same per-campaign and per-sub-ID data, so the weekly pull-and-flag step can run without you opening a spreadsheet — but the block decision still belongs to you until you trust the rules.

What you should not automate is the postback. A broken postback silently turns every site into a "zero-conversion site" and you'll block your best traffic first. Test it after every offer change and every network change.

Frequently asked questions

What is the fastest way to find Taboola sites to block?

Pass Taboola's publisher site token into sub1, record conversions by postback, and group your per-campaign report by sub1. Any site that exceeds your cost tolerance with zero conversions is a block candidate. This gives you spend, clicks and conversions in one table, which Taboola's own reporting can't do on its own.

How many clicks should a site get before I block it?

Divide the amount you're willing to lose testing one site by your average CPC on that campaign. If your CPC is $0.30 and you'll risk $15 per site, that's 50 clicks. A site below that threshold is undecided, not bad — blocking earlier removes traffic that might have converted.

Should I block sites in Taboola or only in my tracker?

Both, but the Taboola exclusion list is what stops the spend. Blocking in your tracker only cleans your reports. Add the domain to Taboola's campaign site exclusions, log the change, and propagate the list to sibling campaigns in the same geo and vertical.

Why do Taboola and my tracker show different click counts for the same site?

Taboola counts the click when its redirect fires; your tracker counts it when the landing page request arrives. Bots, blocked scripts and slow redirects create the gap. A large per-site gap is a signal to check for invalid traffic, not just a reporting quirk.

Can I do this on Adtraxo's free plan?

No. Sub-IDs sub1–sub5 and S2S postback conversions are Pro features. The free plan gives you 10 tracking links, 5,000 recorded clicks a month, one custom domain and basic stats. Pro is $27/month at the founding rate (rising to $47 for new accounts; existing subscribers keep $27) and includes unlimited links, clicks and domains, fraud detection and rules, A/B split testing, country analytics and postbacks.

Ready to see which Taboola sites are actually costing you money? Set up a tracking link, map the site token to sub1, and let the per-site report do the arguing. Start at adtraxo.com.

Track your affiliate campaigns for free

Clicks, conversions, fraud detection — all in one tracker.

Get started free →

Not ready for an account? Get the Tracking Domination Playbook — free, just your email.