Should you gate your demo? What the evidence supports

A gate is a form that must be completed before a viewer can see the demo. Almost every page answering this question is written by a company selling demo software or a rival to it. We sell a product with an email card too, so read this one with the same care. We'll show you what the sources say, what they can't tell you, and how to find out for yourself.

What the public evidence says

This evidence is mostly vendor data, and it doesn't settle the question. We found no independent, controlled study of gated against ungated demos. If you know of one, we'd like to read it.

Why this evidence can't settle it

Look at what the headline number really measures.

It describes winners. A figure about the top 1% tells you what winners did, not what happens when a given team removes a gate. It's like studying only the fastest runners to learn what shoes do.

The ranking may favour open demos. Navattic's report says top performers were picked by traffic and click-through rate. A form at the start stops some viewers, by design, so a gated demo could score lower on that kind of measure even if it produced more sales. The data can't rule that out.

Engagement isn't revenue. More people clicking through a demo is good, but the outcome that matters is qualified buyers. A gate that reduces viewers by half could still be right if the remaining half includes the people who buy. None of the figures above tests that.

It's about one product type. These figures describe interactive demos on one vendor's platform. They may or may not carry over to a film, or to your audience.

The sellers have an interest. That doesn't make them wrong. Navattic's own advice is "It depends". It just means the data was gathered by someone with a stake in the result, and without the controls you'd want.

What's sensible while you lack your own data

Reasoning from first principles gets you most of the way.

Ways to ask for an email without a wall

A gate isn't all or nothing. Here are the common positions, softest first.

• No form. A button at the end, leading to a trial or booking.
• End card. After the film, ask for an email. Viewers who wouldn't have converted lose nothing.
• Dismissible card part-way. A prompt a few seconds in, or at a natural break, that viewers can close.
• Required form part-way. Show the core value, then ask before the rest.
• Required form first. The classic gate.

Programmable covers the first three. A sign-up card can appear when the film ends or a few seconds in, and it shows on your share link and embeds. Visitors can close it and keep watching, so it isn't a hard gate. For that, put a form on your own page, in front of the embed. The card has a consent line, which you should use to say who will email them and how to opt out. Sign-ups can go to your CRM by webhook.

How to test it on your own traffic

This plan compares an open demo with an end card. You can add a third variant if you have plenty of traffic.

1. Write the decision down first. For example: "If qualified demos booked per 1,000 demo viewers isn't at least 20% lower with no card, we'll stay open." This keeps you from reading the results to suit yourself.

2. Choose the right metric. Use something near revenue, such as qualified meetings, trials activated or pipeline created. Don't use emails collected. A gate nearly always collects more emails than no gate, so that metric decides the test before it starts.

3. Make the variants. Turn on the email card in just one version. Change nothing else.

4. Split the traffic at random. In Programmable, publish the film, embed it with the one-line script, and choose Create variant B in the editor's Analytics tab. Variant B starts as a copy, so turn on Collect emails there, then open both versions yourself to check that only B shows the card before you send traffic. The embed splits new visitors 50/50 and each keeps their variant. Judge by CTA clicks: the test also offers sign-ups, but only one version collects them here. A share link or plain iframe isn't split, so for those use your site's testing tool or a small server-side split. Storylane offers A/B testing from its Growth plan, and other demo tools may too. If you can't split at random, alternate whole weeks. That's weaker, because weeks differ, but it's better than comparing a before with an after.

5. Work out how many viewers you need. Evan Miller's rule of thumb is n = 16 × p(1 − p) ÷ d², where p is your current rate and d is the smallest change worth detecting. Read n as the number per variant. The table shows our arithmetic from that formula.

Those counts are demo viewers, not site visitors. At 500 demo views a month and a 2% rate, the test would take about a year, which is too long. Most small sites should choose a higher-rate metric, test a bigger change, or use the decision table above and move on.

6. Set the length and don't peek. Run whole weeks, at least two, so weekdays and weekends are both covered. Miller's article explains why you shouldn't stop as soon as a result looks significant: "Repeated significance testing always increases the rate of false positives." Fix the sample size first and wait for it.

7. Follow the leads downstream. Check analytics for plays, watch-through, CTA clicks and sign-ups for each copy, then match leads to outcomes in your CRM. Give sales cycles time to play out. Decide the final outcome only after that period has passed.

8. Decide, then write it down. Record the numbers, what you changed and when. If the difference is too small to measure, choose the option that's kinder to the viewer.

What goes wrong in these tests

• Counting emails, not buyers. Fake and personal addresses inflate the gated side. Score them by what they do next.
• Mixing audiences. Paid and organic visitors behave differently. Split evenly, and report them separately if you can.
• Changing things mid-test. A new headline or campaign halfway through muddies the result.
• Ignoring sales' workload. More low-intent leads cost rep time. Include that cost.
• Treating one result as permanent. A result holds for that page, audience and season. Retest when things change.

If you do keep a gate

Ask for as little as you can, usually a work email. Tell people what they'll see and how long it runs. Show a preview, such as a short film, above the form. Follow up with something useful, and don't send a rep within minutes of a first look. If you haven't yet chosen a format, read about demo videos and interactive demos first, because the two invite different approaches to asking.

Common questions