Podcast Advertising Metrics: 2026 Benchmarks, Formulas and KPIs

1. Downloads Count Requests, Not Listeners

A download happens when a device pulls an episode file from a server. That pull can be automatic. It can fire while a listener sleeps and an app refreshes a show they stopped opening years ago.

The IAB Tech Lab sets the technical standard that strips bots and duplicate requests out of raw figures. It released Podcast Technical Measurement Guidelines v2.3 on July 21, 2026, open for public comment until August 19, 2026. The update covers audio and video podcasts distributed through open RSS feeds and podcast apps.

The same guidelines explain why the gap survives certification. Most podcast apps still do not confirm playback back to the server. Measurement therefore runs on server side log analysis, so a compliant download stays a request rather than a proven listen.

Nothing here says downloads are useless. They set the outer edge of who could have heard you. Every other metric in this guide narrows that edge toward the people who actually did.

2. The Six Metrics That Matter in 2026

Six numbers decide whether a podcast buy worked. Each one answers a question the download figure cannot, and each has a different source. The table below is the whole framework in one screen.

2026 reference points for each metric

These figures come from published campaign data, not from rate cards. Use them to filter shows and to sanity check results, never as promises.

One number in that table deserves a second look. Magellan AI reports that 2.29% of reached listeners visited the advertiser site, and that 9.74% of those visitors converted to a lead. Multiply the two and you get roughly 0.22% of reached listeners becoming a lead. That is the planning figure most media plans are missing.

Screening shows before you request any of this saves days. MillionPodcasts indexes 3M+ shows, 43.1K of them with a sponsor already on record. Its 17 filters cover audience size, listener demographics and episode length. You then ask only shortlisted shows for completion and rate card data.

3. What Is a Good Episode Completion Rate?

Completion rate is the share of listeners who reach a set point in an episode, usually 80% of runtime. It is the most predictive engagement signal you can request, and first time buyers rarely ask for it.

Think about what the number represents. A listener who reaches the 80% mark of a 45 minute episode has given you 36 minutes of sustained attention. If your ad runs mid-roll, it landed inside that stretch.

How to read a completion rate you are given

These are buying filters, not published industry benchmarks. They exist to tell you when to press harder in a negotiation.

How to request the right number

Ask for completion data on episodes of the same length as the ones you will appear in. A 15 minute episode and a 90 minute episode on one show can behave very differently. Ask for mid-roll retention too, not only the show average.

Some shows cannot produce the figure at all. Treat that as information rather than a dead end. Shows with strong numbers usually offer them without being asked twice.

4. Pre-Roll, Mid-Roll and Post-Roll Compared

Placement decides which listener state your ad meets. Pre-roll plays before the content. Mid-roll plays once the listener has committed time. Post-roll plays after the episode ends.

You will find precise listen-through percentages for each placement quoted across the web. Almost none of them trace back to a published study. Ask the show for its own placement level retention instead, and treat any figure it cannot evidence as an estimate.

Runtime is the missing number under ad load

Ad load is a share of runtime, so runtime decides what that share actually buys. A 10.9% ad load on a 20 minute episode is barely two minutes of airtime. The same percentage on a 70 minute episode runs close to eight.

So we checked our own MillionPodcasts database in August 2026 to see how long shows really run, and whether sponsored shows run longer. Every count below covers podcasts that published an episode in the last 12 months.

Read the last column downward. Sponsors sit on 2.3% of shows under 10 minutes and on 10.3% of shows running 60 to 90 minutes. The rate climbs at almost every step up in length, then eases past 90 minutes.

The skew shows up in the totals as well. Of shows already carrying a sponsor, 44.9% run 40 minutes or longer, against 30.1% across all active shows. Advertisers are concentrated in the long half of the market, and the 20 to 40 minute bucket is where the unsold inventory sits.

Length alone does not earn a sponsor. Longer episodes simply hold more mid-roll inventory, and mid-roll is where direct response money goes. If a 15 minute show quotes you a mid-roll rate, ask exactly where in the runtime that spot lands.

What the published data does support

Podscribe found host read placements deliver a median 0.021% conversion per impression, ahead of producer read at 0.019%. Its Q1 2026 report also found direct show buys produce the strongest visitor and purchase rates. Programmatic buys deliver the lowest cost per acquisition, because CPMs are lower and targeting is broader.

Ad load is the placement factor buyers forget. Podscribe put ad load at 10.9% of episode runtime in Q1 2026, and performance generally falls as that share climbs. Host read spots hold up better than pre-produced ones in crowded episodes.

5. Your CPM Is Not What the Rate Card Says

Once you hold a completion rate, you can work out what you really pay per engaged listener. No rate card does this adjustment for you. Run it yourself on every show you consider.

A $30 CPM on a show with 70% completion costs $43 per 1,000 listeners who stayed for your ad. The same $30 on a 40% completion show costs $75. The media kit shows the first number, and your finance team eventually sees the second.

The formula also reorders your shortlist. A $50 CPM show at 78% completion is a better buy than a $22 CPM show at 44%. Raw CPM comparison hides that completely.

Build the comparison sheet before you negotiate

Four columns are enough. Show name, published CPM, reported completion rate, adjusted CPM. Sort by the fourth column and the ranking rarely matches the second.

Where current CPM rates live

This guide teaches the adjustment, not the going rate. Our podcast advertising CPM rates guide carries current pricing by genre, show size and format. Bring the number it gives you back to the formula above.

6. Set Your CPA Ceiling Before You Negotiate

CPM tells you what impressions cost. Cost per acquisition tells you what a customer cost. Only the second one decides whether the campaign made money.

Your CPA ceiling comes from your margin, not from anyone's rate card. Work it out before you send the first email. Skip that step and the rate card quietly becomes your benchmark.

A $250 product at 55% margin nets $137.50 per sale. At a 2.3 times target return, your CPA ceiling is $60. That single number settles most internal arguments about whether a price is fair.

Project the outcome before you accept a price

Start from engaged listeners rather than downloads. Episode downloads multiplied by completion rate gives you the audience your ad can reach. A show with 10,000 downloads at 70% completion offers 7,000 engaged listeners.

Now apply a response estimate. Magellan AI measured a 2.29% visit rate and a 9.74% visitor to lead rate in Q1 2026. That works out near 0.22% of reached listeners becoming a lead. Against 7,000 engaged listeners, that is roughly 15 leads per episode.

At a $60 CPA ceiling, 15 leads justify about $900 per episode. If the show asks $600, you have headroom to test. If it asks $2,500, the maths only works with a far better conversion rate than the benchmark.

7. How to Calculate Podcast Advertising ROAS

Most advertisers can state what they spent. Far fewer can state what came back. ROAS answers that, but only after the attribution window closes.

A campaign costing $5,000 that attributes $23,500 in revenue returns 4.7 times. Whether that clears your bar depends on your margin, not on any published average. Set the bar with the same margin logic you used for your CPA ceiling.

An Acast study with Dentsu put long term podcast ROAS at 4.9 times. The average across media channels was 3.7 times. Long term is the operative phrase. Measured on a short window, the same campaign reports a fraction of that.

Choose direct or programmatic on the metric you care about

Podscribe found direct show buys deliver the strongest visitor and purchase rates per impression. Programmatic buys produced the lowest cost per acquisition, helped by cheaper inventory and broader targeting. Audience buys reached 25.4% of tracked impressions in Q1 2026, up from 19.3% a year earlier.

So the choice is not quality against junk. It is response rate against efficiency, and your ROAS target decides which one wins.

The 90 day lifetime value check

A $60 CPA looks expensive against a $90 product and sensible against a subscription that retains for 18 months. After your window closes, pull the 90 day value of customers attributed to the campaign. Compare it against your channel average.

If podcast customers are worth 1.4 times your average, a campaign reporting 3.0 times ROAS is really returning 4.2 times. That comparison decides whether podcast joins your permanent media mix.

8. Fill In This Scorecard Before Signing

Campaigns without pre-agreed targets end in interpretation. Results arrive, and the meeting becomes a debate about what the numbers might mean. A scorecard written before launch removes that debate.

Share the finished sheet with everyone who will comment on results. If a metric was not on it before launch, it cannot become the headline KPI afterwards.

Media structure is a separate decision

How many shows and how many episodes is a planning question, not a measurement one. Cumulus Media's 2025 Audioscape found no brand is close to saturating podcast audiences. It recommends putting 5% of digital budgets into the channel. Underspending, not overspending, is the common failure.

Our guide to testing podcast ad campaigns covers how to structure a pilot before you commit real budget.

9. How to Measure Podcast Ad Performance

No pixel fires when a listener hears your ad. There is no link to click at the moment of exposure. Someone hears you on Monday, thinks it over, and buys on Thursday from a different device.

Whether that purchase appears in your data depends entirely on what you built before episode one. Four methods together give you the fullest picture, and each catches conversions the others miss.

• Unique promo codes: one code per show tells you exactly which show drove a sale. Create the code before the host records, because changes afterwards need a re-record. Two word codes are easier to remember than alphanumeric strings.
• Vanity URLs: a show specific slug on your domain tracks clicks and landing page visits. The host needs the exact URL to speak it, so set it up before recording. Send it to a dedicated page so non-buyers still enter your remarketing pool.
• Pixel based attribution: platforms such as Podscribe or Claritas match converted customers against listener data to identify the episodes they heard. This catches the listener who searched your brand and never touched a code.
• Post purchase surveys: one question at checkout asking where the customer heard about you. List podcast as an option alongside show names. It surfaces conversions every automated method missed.

Why a 7 day window hides working campaigns

Podcast conversions do not arrive on a tidy schedule. Magellan AI measures its own response benchmarks on a 30 day lookback from ad exposure, which reflects how slowly listeners act. Close your window at day 7 and most of that response never reaches your report.

Conversions also tend to arrive in waves. One cluster lands in the days right after an episode drops. Another follows from listeners who downloaded early and played late, and a third trickles in as new listeners find the back catalogue.

Measuring awareness when there is no link

Some campaigns exist to shift perception rather than close sales. Brand lift is measured by surveying listeners of your shows. Compare their answers against a control group that never heard the ads.

Nielsen's brand impact norms give you the yardstick for that comparison.

Nielsen also found 70% of exposed listeners recalled the advertised brand. Sounds Profitable, surveying 5,005 US adults with Signal Hill Insights, recorded an 86% ad recall rate among the most active podcast users. That was the highest recall of any ad supported medium in the study.

A cheaper signal sits in your own analytics. Branded search volume in the days after an episode drops tells you whether attention moved, and it costs nothing to check.

10. What Your Data Says at 30 and 60 Days

Campaign data arrives in stages, so one snapshot date gives you a misleading picture. Two fixed checkpoints give you enough signal to decide without missing the renewal window.

The 30 day read

At day 30 you have direction, not a verdict. Review promo redemptions and vanity URL traffic. Then pull branded search for the seven days after each episode and compare it with your previous four week baseline.

A branded search rise that holds for several days is a real signal. It often appears while code redemptions still look flat, because most listeners never use the code.

The 60 day read

At day 60 the direct response picture is complete. Pull fully loaded CPA across every attribution method and set it against your ceiling. Calculate ROAS over the full window, then read the trend across episodes.

Before you call a flat result a failure, check the size of the buy. Cumulus Media's Audioscape work found most brands are nowhere near saturating podcast audiences. A campaign too light to register is a different problem from a campaign in the wrong place.

If you are buying for direct response, pick one well matched show and measure it properly for 60 days. One show with clean attribution teaches you more than five shows with none.

If you are buying for awareness, agree the lift study with a measurement partner before launch. Then hold the same shows long enough to survey the audience that heard you.

Either way the first hour of work is identical. Open a sheet and write your CPA ceiling and ROAS target. List the four attribution methods that will be live on day one. Do that before you request a single rate card, and every number that follows means something.

11. Podcast Advertising Metrics FAQ

What podcast advertising metrics matter most?

Six numbers decide whether a campaign worked: episode completion rate, adjusted CPM, ad load, cost per acquisition, ROAS and attribution coverage. Downloads sit outside that list because they count file requests. Fix your CPA ceiling and ROAS target before a rate card reaches you.

How do you measure podcast advertising performance?

Run four attribution methods at once. Use a unique promo code per show, a vanity URL, pixel based attribution and a post purchase survey. Each one catches conversions the others miss. Put all four live before the first episode airs, because data you did not capture cannot be recovered later.

What is a good episode completion rate for podcast advertising?

Above 70 percent is strong for direct response, because a listener who finishes the episode almost certainly heard a mid-roll. Between 50 and 65 percent still works for awareness goals. Below 50 percent, ask why before you commit budget.

How long should a podcast attribution window be?

Use 30 to 60 days. Magellan AI measures its own response benchmarks on a 30 day lookback from ad exposure. Listeners routinely convert days later, often on another device. A 7 day window makes working campaigns look like failures.

How is ROAS calculated for podcast advertising?

Divide revenue attributed to the campaign by total ad spend, using the same window you use for cost per acquisition. An Acast study with Dentsu put long term podcast ROAS at 4.9 times, against a 3.7 times average across media channels.

How long is the average podcast episode?

MillionPodcasts data from August 2026 shows the 20 to 40 minute bucket is the single largest, at 210.4K shows active in the last 12 months. Shows carrying a sponsor sit longer, with 44.9 percent of them running 40 minutes or more against 30.1 percent of all shows.

How do you measure podcast success without a promo code?

Watch branded search volume in the days after each episode drops, and compare it with your baseline. Add pixel based attribution to catch listeners who search your brand instead of typing a code. For awareness campaigns, a survey against a control group is the only reliable read.

References