How do you get UGC videos for a SaaS product with no physical item?

Three routes work. Screen recording with a face cam is the closest equivalent to an unboxing and is what most creators produce for software. Founder explainers need no creator at all. Generated video skips the creator entirely and takes your own screenshots or footage as the product asset. All three are legitimate; they differ in what each additional video costs you, since creator routes charge per video while generated ones charge per second of output.

Why is UGC harder for software than for physical products?

Because the thing you would normally ship does not exist. With a physical product you mail a sample and the creator films it. With software you have to provision access instead: a working account, a demo environment, believable sample data, and a briefing so the creator does not record the wrong screen or expose someone's real data. That work repeats for every creator, and unlike a sample it cannot be done by a warehouse.

How much do UGC videos cost for tech and app brands?

Managed UGC agencies serving tech and app brands quote roughly $150 to $300 per finished video, plus your own time provisioning access and reviewing footage. Marketplace creators run cheaper per video but need more briefing for software. The number that matters is not the unit price but the per-video slope: every additional creative costs another full unit, which is what makes wide creative testing expensive on this route.

Which product categories cannot use creator UGC at all?

Anything where the sample or access cannot be handed over at the scale testing requires. Software and apps have no physical sample. Subscriptions and courses have nothing to unbox. Oversized or heavy goods cost more to ship as samples than the creator fee. High-ticket items cannot be given away in volume. Pre-launch products have no inventory yet. In all of these the creator route is not expensive, it is unavailable.

Do screen recordings actually work as UGC ads?

Yes, and they are the standard format for software. A screen-in-hand demo, a founder explainer, or a problem-then-solution walkthrough all perform as UGC on short video. The constraint is not whether the format works but how many of them you can afford: each one still costs a full production unit, so most software brands run a handful of creatives rather than the dozens that creative testing actually calls for.

Can you make UGC style video ads without hiring a creator?

Yes. Generated video takes your own assets, a character, and a script, and produces the same short-form shapes creators film. What it removes is the per-creator overhead: no sample, no account provisioning, no NDA, no scheduling. What it does not remove is the need to know what to say, which is why starting from the structure of a video that already won beats starting from a blank prompt.

How many creatives should a SaaS brand test?

More than most run, and the limit is usually budget for production rather than judgment about what to test. Creative is the variable with the widest spread in paid performance, so the number of distinct angles you can put in front of an audience matters more than the polish of any single one. When a creative costs single-digit dollars instead of a few hundred, the sensible number goes up sharply.

UGC Ads for SaaS and Apps: What to Do When There Is Nothing to Ship

Creator UGC has a hidden assumption baked into it: that there is a physical thing you can put in a box. Run 100 videos and you ship 100 samples. That math is annoying for a physical product. For software it does not apply at all, because there is nothing to mail.

That does not mean UGC is unavailable to you. It means the cost moved somewhere less obvious, and most guides never name it.

The sample is replaced by access, and access is worse

With a physical product, the handover is a package. With software you have to provision the creator into your product:

a working account, usually a comped one
a demo environment, because you cannot let a stranger record inside live customer data
believable sample data, since an empty dashboard films terribly
a briefing so they record the right screens, in the right order, without exposing anything they should not

That is not a one-time setup. It repeats for every creator you work with, and unlike shipping a sample, nobody can do it for you. A warehouse can mail a hundred boxes. Only your team can provision a hundred accounts and check a hundred recordings.

Managed UGC agencies serving tech and app brands quote roughly $150 to $300 per finished video. Marketplace creators come in lower, but software needs more briefing than a skincare product does, so the gap narrows once your own hours are counted.

The number that matters is the slope, not the price

A single $200 video is not a problem. The problem is what happens when you want twelve.

Creative is the widest-spread variable in paid performance. Finding the angle that works is a search problem, and search problems are solved by trying more candidates. So the question is never "can I afford one video" but "what does creative number twenty cost me".

On a creator route, video twenty costs the same as video one. The slope is flat and high. That is why most software brands end up running four or five creatives and calling it a test, when the honest version of that test needs several times more.

Where this stops being a cost and becomes a wall

For some categories the creator route is not expensive. It is unavailable.

Category	Why the handover fails
SaaS and apps	No physical sample exists. Access has to be provisioned per creator
Subscriptions	Nothing to unbox; the value shows up over weeks, not in a 15 second clip
Courses and info products	Same, plus the content is the product, so you are handing over the asset itself
Oversized or heavy goods	Sample freight costs more than the creator fee, and some items cannot be shipped at all
High-ticket items	You cannot give away twenty units to test twenty angles
Pre-launch products	There is no inventory yet, and launch content is needed before there is

If your product is on this list, the standard UGC playbook was never written for you. That is worth naming plainly, because a lot of advice aimed at physical-product sellers gets recycled at software brands where the first step is impossible.

Screen recording works. Be honest about what it costs.

None of the above means screen recordings are a poor format. They are the right format for software, and they perform:

screen-in-hand demo, phone in frame, thumb moving through the flow
founder explainer, which needs no creator at all and often outperforms a hired one
problem, then solution, where the first three seconds show the annoying manual version
"why I switched", the comparison format, which converts well when the switch is specific

Lead with the outcome rather than the feature list. "I got three hours of my week back" lands harder than "automated workflows". These are craft points, and they matter.

They just do not change the slope. Every one of those videos is still a production unit with a person, a schedule, and a review cycle attached.

What changes when the video is generated

Generated video removes the handover entirely. There is no sample because there is no creator, and no account to provision because the product footage comes from assets you already own: your screenshots, your recorded flows, your product images.

What that removes:

no sample, no shipping, no freight on heavy goods
no account provisioning, no demo environment, no NDA
no scheduling, so the turnaround is minutes rather than the two to three weeks a creator round takes
no sunk cost, which matters more than it sounds. A creative that cost $200 is one you will keep running to justify. A creative that cost a few dollars gets switched off the moment it underperforms, and that decision is where the real money is

What it does not remove is the hard part: knowing what to say. A blank prompt is a bad starting point, and it is why so much generated video reads as filler. Starting from the structure of a video that already won is a different exercise. You are not inventing a hook, you are working out why one worked and rebuilding that shape around your own product. We wrote about that distinction in formula riffing versus script-to-video, and about the comparison with hired creators in AI UGC versus real creators.

If you sell a physical product but do not have a sample in hand yet, that is a different situation with a different answer: see making TikTok ads without the product, which is written for affiliates working from supplier assets rather than for brands whose category has no sample at all.

What this does not fix

Two honest limits, because the category is full of claims that skip them.

It does not make a weak offer convert. A video can only get the right person to look. If the product does not solve a visible problem or the pricing page loses them, more creatives will find that out faster and cheaper, not fix it. Views and orders are separated by everything that happens after the tap.

It does not remove judgment. Which angle, which audience, which claim to lead with: those decisions still belong to you. What changes is that you can afford to be wrong about them repeatedly, which is the only way anyone finds out.

The practical version

If you sell software, a subscription, a course, or anything you cannot put in a box:

Stop budgeting for samples. That line does not exist for you, and the guides that assume it are not describing your situation.
Count provisioning as production cost. Account setup, demo data, and review time are real hours. They are the reason your creator route feels slower than the quoted turnaround.
Decide how many angles you actually want to test, then check whether your creative route can produce that many. If it cannot, the route is the constraint, not your ideas.
Keep the craft. Outcome-first scripts, a real first three seconds, a specific comparison. Those rules do not change based on how the video was made.

You can start with one video free and see what the shape looks like on your own product before deciding anything.