Someone in the mastermind says their VSL converts at 3%. Yours is at 1.2%. For the rest of the call you are quietly wondering whether to rewrite the hook, fire the copywriter, or kill the offer.
Nobody asked them the obvious question. 3% of what? Visitors, plays, or the people who stayed to the pitch? On the same funnel, those numbers can be ten times apart.
For cold paid traffic, most VSL funnels convert roughly 0.5 to 2% of page visitors into buyers, and warm traffic from email or retargeting often 2 to 6%. None of these public ranges comes from an audited dataset, and they only compare if you use the same denominator: per visitor, per play, or per viewer who reached the pitch.
The benchmark that counts is your own funnel last month, step by step.
The ranges, with the caveats up front
There is no public, audited dataset of VSL purchase rates. What circulates is a handful of published estimates and a lot of consultant anecdotes. The most careful one we found is a benchmark page from Daily Intel, which at least defines its rate as completed sales divided by unique VSL page visits. It also labels its own ranges "directional estimates".
| Published range | What it divides by | How solid it is |
|---|---|---|
| Cold paid social: 0.30 to 1.80% | Unique VSL page visits | One publisher's estimate, no dataset |
| Retargeting: 1.80 to 5.50% | Unique VSL page visits | One publisher's estimate, no dataset |
| Email or SMS list: 2.00 to 6.00% | Unique VSL page visits | One publisher's estimate, no dataset |
| ClickBank: favour offers at 1% or more | Hops (affiliate link clicks) | Vendor guidance for picking offers, not an average |
| Landing pages: 6.6% median | Visitors, any goal | Large dataset, but not purchases and not VSLs |
The ClickBank line comes from its own advice to affiliates, which is to prefer offers with a conversion rate of 1% or more and an EPC of $1 or more. It is a filter for choosing offers, not a measured market average. The landing page median gets its own section below, because it is the number most often quoted by mistake.
Use the ranges to spot a funnel that is badly broken. Do not use them to decide whether yours is good.
Every conversion rate has a hidden denominator
A conversion rate is a fraction. The top is almost always the same: orders. The bottom is where everyone quietly disagrees, and the bottom decides the number.
Take the worked example from our guide to VSLs: a 16-minute video, a $97 offer, 10,000 plays in a week, 140 orders. Add the page visits behind those plays and you get four honest conversion rates for one funnel.
| Denominator | Count | Orders | Rate |
|---|---|---|---|
| Page visitors | 20,000 | 140 | 0.7% |
| Genuine plays | 10,000 | 140 | 1.4% |
| Still watching past 5 seconds | 7,200 | 140 | 1.9% |
| Reached the pitch | 1,400 | 140 | 10% |
An illustration, not customer data. The visitor row assumes half of page visits became plays.
Same video, same week, same 140 orders. The mastermind "3%" could sit anywhere on that table.
Per visitor
Orders divided by people who landed on the page. This is what most blog posts mean, and what affiliate networks report. ClickBank's hop conversion rate is initial sales divided by hops, and since May 2024 it is calculated on unique hops rather than raw clicks. It judges the page, the player and the traffic together, which is useful for buying media and useless for diagnosing the video.
Per play
Orders divided by plays. In TrackPlay this is the Conversion Rate card: plays that ended in a sale, with the sale posted back by your cart onto the session that produced it. It takes the page out of the question and asks what the video is worth once someone is actually watching. Its money twin is Revenue per Play, $1.36 in the example.
Per viewer who reached the pitch
Orders divided by the viewers still there when the offer was revealed. This is the number that judges the pitch itself, the offer, price and close, separated from everything that happened before it. In the example it is 10%, which says the pitch is doing its job and the leak is upstream.
Why "a play" is not one number either
Per play only compares if two hosts count plays the same way, and they do not. Some count a muted autoplay impression as a play. Some count every reload. Some count bots. A host that counts autoplay impressions reports more plays for the same sales, so the same funnel shows a lower per-play rate.
TrackPlay counts a play as one session in which the viewer genuinely played: pressed play, turned the sound on, resumed, or continued. A muted autoplay nobody watched is not a play. A bot is not a play. Reloading mid-session is not another play. The full definition is public in the Honest Play Standard, and billing and every report use the same number.
The funnel of ratios: where the number is set
One conversion rate is really five smaller ratios multiplied together. Each has its own metric, and each is set by a different part of the funnel. That is the useful way to read it: not "is 1.4% good?" but "which step is losing the most people?"
- Visit to play. Play Rate, page loads that became genuine plays. Set by the page, the poster, the player's first frame and how autoplay is handled.
- Play to hook. Hook Rate, the share still watching past five seconds. Set by the opening lines.
- Hook to pitch. Hold Rate, the share that reached the midpoint, or the Pitch reached step when a pitch time is set. Set by the story, the mechanism and the proof.
- Pitch to click. Click Rate on the buy button. Set by the offer, the price and when the button appears.
- Click to order. The sale your cart posts back onto the session. Set by the checkout.
The funnel report lays these out from page view to sale, with the drop at every step, and the metrics catalogue prints the formula on each card. In the example, 2,800 of 10,000 viewers left inside the hook. That is the biggest single loss in the chain, and no amount of pitch rewriting reaches those people.
For outside context on the first step only: in Wistia's 2026 data, about 24% of homepage visitors and 15% of product page visitors press play. Those are general marketing videos sitting on pages built for something else. A VSL page is built around the video, so its play rate should sit well above those. If yours looks like a homepage number, start there.
What actually moves the number
Five things decide most of the gap between your rate and the one someone quoted at you. Only some of them are about the video.
Traffic temperature
Warm traffic converts better than cold. Every source agrees on the direction and nobody has audited the size. Daily Intel's page says warm audiences often convert at two to four times the rate of cold ones. The closest large dataset is landing pages, not VSLs: in Unbounce's benchmark report, email traffic converted at 19.3% against 12% for paid social.
So never blend them. A week where retargeting spend went up will look like the VSL improved, when nothing about the video changed.
Price point
A higher price means fewer buyers per hundred visitors. Daily Intel's price bands, again estimates, put cold traffic at 0.45 to 1.50% for a $27 to $97 offer and 0.08 to 0.45% for $497 to $2,000. On those numbers, 0.4% is poor for the first and healthy for the second. Comparing offers across prices on rate alone will always push you toward the cheapest one. Compare them on revenue per play or EPC instead.
Niche and offer
The market, the proof and the guarantee move the median before the video does anything. In Unbounce's landing page data, industry medians alone run from 3.8% to 12.3%, more than three times apart, with the same kind of page and the same way of counting.
Device and delivery
Unbounce found that mobile brings nearly 5 times the visitors while desktop converts 8% better. So a rate that dips when your mobile share rises is not necessarily a worse video. Compare devices apart before judging. Delivery matters too. Error Rate and Buffer Rate sit in the same metrics catalogue as Conversion Rate, so a player that failed to start can be told apart from a hook that failed to hold.
Definition drift
The quiet one. What your host counts as a play, and what your cart counts as a sale: front end only or upsells too, before refunds or after. Change either and the rate moves with no change in the business. Write the definitions next to the number, every time you report it.
VSL vs landing page conversion rates
The most quoted "good conversion rate" is the Unbounce landing page median of 6.6%, drawn from 41,000 pages and 57 million conversions. It is solid data. It is also not a VSL benchmark.
| Landing pages (Unbounce) | Median conversion rate |
|---|---|
| All industries | 6.6% |
| Ecommerce | 4.2% |
| SaaS | 3.8% |
| Financial services | 8.3% |
| Education | 8.4% |
Unbounce counts any conversion goal, a form, a signup, a click or a purchase, per visitor. Many of those pages ask for an email address. A VSL is usually asking for a credit card. Put a purchase rate next to an opt-in rate and the VSL will always look like it is failing. For how the two formats differ beyond the numbers, see VSL vs. traditional landing page.
Benchmark against your own baseline
Every public number above is either an estimate, a different kind of page, or a different denominator. The only benchmark with your traffic, your price and your definitions in it is your own funnel last month. TrackPlay does not publish a cross-account average for the same reason. Every card compares against your own previous period instead.
- Pick one denominator per report and write it down. Per visitor for media buying, per play for the video, per pitch viewer for the offer.
- Track each step week over week. Every metric card shows a versus-last-period change and a sparkline, so a drop is visible at the step where it started.
- Change one step at a time. A new hook and a new price in the same week leave you with a moved number and no idea why.
- Test on posted-back orders. A split test in TrackPlay scores each arm on a rate per unique play, fed by the cart's postback, so an arm cannot win by getting more traffic.
- Compare sources on money. The Sources report shows conversion rate, EPC and revenue per source, and anything it cannot place is labelled Unknown rather than guessed as Direct.
Across offers at different prices, finish on Revenue per Play. A $97 offer at 1.4% of plays earns $1.36 a play. That number talks straight to your ad account: if a play costs more than that in media, the rate was never the problem.
The one-paragraph version
Published VSL ranges put cold traffic at roughly 0.5 to 2% of visitors and warm traffic higher, and none of them is audited. Before comparing any rate, ask what it was divided by: visitors, plays and pitch viewers can be ten times apart on one funnel. Break your rate into its steps, fix the one losing the most people, and judge offers on revenue per play. If you want to see those steps on your own VSL, the free plan gives you 3,000 plays a month, free forever, with no card.