SaaS Video Ads: Which Format Fits Each Funnel Stage

A widely repeated piece of advice says the best SaaS video ad is the one that shows your product. Lead with the dashboard, the integrations, the moment something automates itself, and the quality of the software will carry the campaign. It sounds reasonable, and it is wrong often enough to be expensive. The same demo that converts a warm trial-page visitor will drain a cold prospecting budget, because the two viewers are not asking the same question. Strong SaaS video ads are not built around the best footage. They are built around what a viewer at a given funnel stage is ready to hear.

That distinction explains most of the gap between campaigns that scale and campaigns that stall. Reviewed in isolation, a creative only invites one question: is it a good ad? The more useful question is where it belongs.

Why the funnel stage chooses the format

Picture the same person at three moments. First, they have never heard of you and are not looking for you, so your ad is an interruption. Second, they recognize the category and have seen your name once or twice, so they are half-deciding. Third, they are comparing you against one alternative and want a reason to act today. Marketers label these stages top, middle, and bottom of funnel, or TOFU, MOFU, and BOFU, but the labels matter less than the shift behind them:

• TOFU (cold prospecting): earn attention and produce a problem-aware click. The viewer cares about their own situation, not your feature set.
• MOFU (consideration and first retargeting): build belief. The viewer knows the category is real and needs evidence that your version works.
• BOFU (decision and warm retargeting): clear the last objection and trigger the action. Intent is high, so the viewer tolerates more friction in the ask.

Run a format at the wrong stage and nothing breaks loudly. The auction still serves the ad, the dashboard reports clicks, the budget spends. You simply lose money on a delay, the failure mode hardest to notice in a weekly report.

Five SaaS video ads, and where each one belongs

Most B2B video ads that perform are a version of one of five formats. The useful way to hold them is not by their look but by the job each one does and the stage that job fits.

Problem-agitate (TOFU)

This format names a pain the viewer feels regularly, makes it concrete, and points toward the category as the way out. The product barely appears, because the goal is recognition, the small jolt of a viewer thinking the ad is describing their week. Open on the pain in the first couple of seconds, show one specific consequence, then offer a glimpse of the better state. Keep the ask light, a guide or a free tool, because a cold viewer will not commit to a trial off one video. You will run this format in the highest volume, since cold prospecting is where the impressions are and where your scroll-stopping hooks earn the most from rapid testing.

ROI or stat lead (TOFU and MOFU)

Here the hook is a number that reframes the cost of doing nothing, such as "most onboarding flows lose half their signups before activation." Use numbers you can defend, drawn from your own data or a widely cited benchmark you present as a benchmark, never an invented statistic. A fabricated "73 percent of teams" line is both a credibility problem and, on Meta and LinkedIn, a policy problem. The format flexes across stages: a sharp number interrupts the scroll for a cold viewer and reads as proof for a warmer one.

Social proof (MOFU)

This format leans on evidence that people like the viewer already trust you: a customer outcome, a usage milestone, a wall of recognizable logos, or a short quote on screen. The job is belief rather than awareness, which is why it belongs in the middle. By MOFU the viewer accepts that the category exists, so the ad answers the next question: does this work for someone in my situation? Specificity carries it: a named role with a real before-and-after will outperform "loved by thousands," and the question of how much social proof to put on screen matters as much as having any. One firm boundary: do not manufacture testimonials, names, or screenshots. If you have no proof yet, keep running problem and ROI ads until you earn some, or look at how to build a testimonial ad before you have customers without inventing one.

Feature highlight (MOFU and BOFU)

This format takes one feature, not the whole product, and shows the specific job it does and the friction it removes. A feature highlight that parades five capabilities is a brochure and converts like one. Name the friction, show the single feature resolving it, state the payoff. "Reports that build themselves" lands harder than "powerful analytics suite." It works for warm audiences who know you exist and are deciding whether you do the one thing they came for.

Demo clip (BOFU)

This format shows the product doing real work, start to finish, on an actual screen. It is the least persuasive ad for a cold viewer and the most persuasive for a warm one, because by BOFU the viewer is explicitly asking to see how it works. When the product is simple enough, the explainer format can deliver the same understanding in a tighter package. Keep it disciplined even at the bottom: one workflow, the result, the call to action. A two-minute walkthrough belongs on the landing page; the ad confirms a decision, it does not train a new user.

What the cost difference actually looks like

One rule sits underneath the whole map: the colder the audience, the less your product should appear on screen and the lighter the ask should be. The reason is economic. Across most SaaS categories, cost per trial on a cold audience runs meaningfully higher than on warm retargeting, often two to three times higher, because cold traffic carries no prior intent. A B2B SaaS click on Meta or LinkedIn is rarely cheap, and LinkedIn cost per click frequently lands several times above the same offer on a broad audience elsewhere, which is why the B2B LinkedIn playbook treats stage discipline as a budget question. Show a demo to people who do not have the problem yet and you pay top-of-funnel prices for a bottom-of-funnel message. The fix is not a better demo. It is moving the demo to the audience that asked for it and putting the problem-led ad where the cold spend goes.

Worked SaaS ad examples for two categories

Two short scripts show the same logic across different products, as illustrations of structure rather than benchmark claims.

A cold problem-agitate ad for invoicing software might run: "Freelancers still chase late invoices one awkward email at a time. Every month, that is real income sitting unpaid. There is a way to make the follow-ups send themselves on schedule. Grab the free chasing-script template, link below." Around 35 words, near 13 seconds at a normal read, asking for nothing risky.

A warm feature-highlight ad for a customer-support tool might run: "Still copying the same answer into ten tickets a day? Watch this reply write itself from your help center, then send. Works with the inbox you already use, not a separate app. Start a free trial." It shows one feature doing one job and asks for the trial, because the viewer knows the category. Notice that the demo footage appears only in the warm script. Both come from a brief using the same logic as a fuller ad-script framework, and both assume readable captions on screen for a muted feed.

Sequencing the formats across a live campaign

The five formats are not a menu you pick one item from. They are a path the same viewer travels, and the strongest campaigns move people along it. In the cold phase, launch several problem-agitate and ROI hooks at once and vary mainly the opening line, since the winner tells you which pain your market feels. Retarget the engagers with social proof and feature highlights, served to anyone who watched past the halfway mark or clicked, because they have context and need belief. The broader question of retargeting creative is really about what to show people who already bounced. Finally, send a demo clip and a deciding-feature highlight to trial-page visitors who did not convert, and ask for the action directly.

The cold hook stays the highest-leverage variable in that sequence: most variance, cheapest to test. The limit has usually been production. Hand-editing five hooks plus three retargeting formats is most of a working week, which is why many small teams ship one ad per stage and never learn which pain or proof moves the number.

Where this rule of thumb bends

Mapping formats to stages is a strong default, not a physical law. Some tools are simple enough that the demo is the hook, so a clean walkthrough works cold because seeing it instantly explains the value. Some categories are so pain-aware that you can skip agitation and lead with proof. Treat the map as a starting grid and let your own retention and cost-per-action data override any cell it gets wrong for you.

How many variants do I need per stage?

A sensible floor is three to five genuinely different creatives per stage: varied hooks at the top, varied proof in the middle, one or two demo cuts at the bottom. Fewer than that and you cannot read what is working. Near-identical recuts do not help either, since they split the budget across versions that test the same idea. The variants that earn their place change the angle, not the font.

Sources

1. AdAmigo.ai — Meta Ads Benchmarks by Funnel Stage (2026): Prospecting vs Retargeting vs Retention
2. Aimers — Facebook Ads Benchmarks for B2B SaaS: CTR, CPC, CPL
3. Postiv AI — LinkedIn Advertising Costs 2026: CPC, CPM, CPL & ROAS Benchmarks by Industry
4. Closely — LinkedIn Ad Benchmarks 2025: CPC, CPM, and CTR by Industry
5. WordStream — Facebook Ads Benchmarks 2025
6. The Voice Realm — Voice Over Script Calculator (words-per-minute timing)
7. JL Recording Studios — Word Count and Commercial Script Timing
8. Meta Transparency Center — Advertising Standards: Misinformation
9. LinkedIn — Advertising Policies

If production cost is what keeps you from running the right format at each stage, that is the gap Aitachyon closes. It turns one brief into a problem-agitate ad, a feature highlight, and a demo clip quickly enough to run all three across the funnel. See how it fits your stack.