LinkedIn is the reader's test here, because the files are YouTube uploads with no placement data; the hook strategy is what the first 15 seconds do. Across 179 SaaS teasers of one minute or less from 150 software companies, 128 of them uploaded in 2026, 111 speak a line by 0:15 and the voice starts a median 0.3 seconds in. 81 of the 111 first lines open on something other than the product, and the name follows at a median 0:07. Companies in the set include Swimlane, Lyzr, SuperOffice, Global Payments, etc.
ADVIDS Research · 179 SaaS teasers audited · 150 companies · 111 first lines hand-coded · 39 teasers with a full visual record · published 14 September 2026 · 16 min read
Top takeaways
A voice speaks by 0:15 in most SaaS teasers
111 of 179 SaaS teasers of one minute or less, uploaded 6 May 2022 to 22 July 2026, speak a line by 0:15; the other 68 run music with no words, a sung vocal or a later voice, and all 7 Anthropic and OpenAI teasers are music only.
On-screen text reaches the frame before the voice
On-screen text appears by 0:15 in 37 of the 39 visually coded SaaS teasers, at a median 0:02.5, and reaches the screen before the first spoken word in 26; Make, Lyzr and Cegid are among them.
The voice starts inside half a second
In the 111 SaaS teasers that speak by 0:15, the voice starts a median 0.3 seconds in; 69 speak inside half a second, 83 inside one second, and 5 wait past three: Airia, Bluebeam, Papaya Global, Zendesk, etc.
Most first lines open on something other than the product
81 of 111 first spoken lines in SaaS teasers open on something other than the product: problem or pain leads with 25, then the name with 19, a bold claim with 16 and a question with 14; Nooks alone opens on a statistic.
The product name arrives by 0:15, after the first line
77 of 111 speaking SaaS teasers say the product or company name by 0:15, but 32 say it in the first line; the name lands at a median 0:07, and Cohesity waits until 0:38.
Speaking teasers fit a median 33 words into 15 seconds
SaaS teasers that speak by 0:15 fit a median 33 words into the first 15 seconds, 2.2 a second; 73 of 111 speak 30 or more, and 7 speak fewer than 20: Bluebeam, Cohesity, Corpay, Microsoft, etc.
Questions this report answers
How should a SaaS teaser video open in the first 15 seconds?
In 179 SaaS teasers of one minute or less uploaded 6 May 2022 to 22 July 2026, 111 speak by 0:15, the voice starts a median 0.3 seconds in, and 81 of 111 opening sentences are about something other than the product. The name follows at a median 0:07, in teasers from Swimlane, Lyzr and SuperOffice among others.
When should a short SaaS video ad say the product name?
Most speaking SaaS teasers name the product inside the first 15 seconds but not in the opening sentence: 77 of 111 name it by 0:15, at a median 0:07, while 32 name it in the first line. Epicor names itself at 0:00, Cohesity at 0:38, and 13 never say it.
Do short SaaS video ads use a voiceover or music only?
118 of 179 SaaS teasers of one minute or less carry a voiceover and 61 run music only, by the library's audio analysis. All 7 teasers from Anthropic and OpenAI run music only and put their words on screen instead.
How many words fit in the first 15 seconds of a SaaS teaser?
Speaking SaaS teasers carry a median 33 words before 0:15, 2.2 words a second, with the middle half between 26 and 38. Corpay speaks 9 and Paycom 54.
Do SaaS video ads open with a question?
Rarely: 14 of 111 first spoken lines in SaaS teasers are questions, including Swimlane, Zapier and NeuroBlock. Problem or pain opens 25, the product's name 19 and a bold claim 16.
Does this analysis show what works on LinkedIn?
No: the 179 SaaS teasers are YouTube uploads, and the library records no placement, views or clicks. The report measures what the first 15 seconds contain; whether an opening holds in a LinkedIn feed is the reader's test to run with their own campaign data.
How was this analysis done?
ADVIDS Research screened 343 library videos running 0:15 to 1:00 against a rule written before intake, and 179 SaaS teasers from 150 companies qualified. Speech timing came from speech segments and word-timed captions, first lines were hand-coded, and on-screen text came from 39 visually coded teasers.
How to read this report
343 library videos running 0:15 to 1:00 were screened against a rule written before intake, and 179 qualified as short-form promotional videos for a SaaS product, from 150 companies, uploaded to YouTube 6 May 2022 to 22 July 2026. A teaser is one such video; the hook strategy is what its first 15 seconds do.
"111 of 179" is a count of teasers over its base: all 179, the 111 that speak a line by 0:15, or the 39 with a full visual record. Each beat states its base. Times are medians in real seconds, "by 0:15" means before 15.0 seconds, and a pattern is mapped when it holds for at least half its base.
Beat 01 · Sound · median at 0:00 · base: all 179 SaaS teasers
1. A voice speaks by 0:15 in most SaaS teasers
111 of 179 SaaS teasers of one minute or less speak a line by 0:15, 62.0 per cent. The other 68 split three ways: 58 carry music with no words at all, 8 carry a sung track or an incidental vocal, and 2 hold the voice past 0:15. The split follows the library's audio analysis: every teaser coded voiceover carries timed speech, and none coded music only speaks a line. Without Anthropic and OpenAI the speaking share is 64.5 per cent. Two teasers in three hand the first 15 seconds to a voice; one in three hands them to picture and type.
All 7 teasers from the two model labs, Anthropic and OpenAI, run music only. Companies whose teasers speak a line by 0:15 include Bloomfire, NiCE, Cint, Global Payments, etc.
HubSpot says nothing for the whole half-minute. Its June 2026 One Platform teaser runs 0:30 of kinetic type with music only: a dark card, a number ticker from 0:04, the word INSIGHT at 0:09 and a chart line reading drives results. Built this way, every word of the hook is on screen.
Source: ADVIDS Research coded set, 179 SaaS teasers from 150 companies, uploaded 6 May 2022 to 22 July 2026
The four rows add to the whole set: a teaser either hands the opening to a voice or to picture and type.
Beat 02 · On screen · median at 0:02.5 · base: the 39 visually coded SaaS teasers
2. On-screen text reaches the frame before the voice
On-screen text appears by 0:15 in 37 of the 39 SaaS teasers that carry a full visual record, and the first line of it lands at a median 0:02.5. In 26 of 39 that text reaches the screen before the first spoken word: all 20 music-only teasers, where type is the only carrier of words, and 6 of the 19 that speak by 0:15. The opening frame itself carries no words in 22 of 39: 17 open on a blank card or abstract ground that holds a median 4.0 seconds, 9 open on written type, 10 on people, 2 on the product interface and 1 on a logo card. In 7 of the 7 speaking teasers that open on a blank card, the voice starts before the card ends. The first thing a viewer can read arrives within seconds, and in two teasers in three it arrives before anything is said.
Companies whose on-screen text arrives before the voice include Anthropic, OpenAI, Alteryx, Make, etc. The 2 with no on-screen text by 0:15 are Proofpoint and Showpad. The opening segment carries no text in 22 of 39, and in 22 of 33 without Anthropic and OpenAI, a shift of 10.3 points, so that row is support and not a finding.
Proofpoint puts no words on screen in its first 15 seconds. Its February 2026 spot runs 0:30 of live action, a man walking through an office past a fire that breaks out unnoticed, while the voice speaks 33 words from the first second. Here the voice does all of the reading.
Source: ADVIDS Research coded set, 179 SaaS teasers from 150 companies, uploaded 6 May 2022 to 22 July 2026
17 of the 39 first frames carry no words; the first readable line follows a median 0:02.5 in.
Beat 03 · First word · median at 0:00.3 · base: the 111 SaaS teasers that speak by 0:15
3. The voice starts inside half a second
In the 111 SaaS teasers that speak by 0:15, the first speech segment starts a median 0.3 seconds in, with the middle half between 0.1 and 1.1 seconds. 69 start inside half a second, 83 inside one second and 106 inside three; 5 wait longer, the latest at 5.5 seconds. Timing comes from word-timed captions for 25 of the 111 and from speech segments for the other 86; on the 81 library videos that carry both, the two placements of the first word differ by a median 0.02 seconds. In a spoken teaser the words do not wait for the picture.
Companies whose voice starts inside half a second include ElevenLabs, Shoplazza, Akeneo, SuperOffice, etc.
Bluebeam holds its voice for five and a half seconds. Its May 2026 Bluebeam Max teaser, 0:39 long, starts speaking at 0:05.5 with "In the wild west of construction, every second counts", the latest start in the set, and still names the product at 0:17.
Source: ADVIDS Research coded set, 179 SaaS teasers from 150 companies, uploaded 6 May 2022 to 22 July 2026
Beat 04 · First line · median at 0:00.3 · base: the 111 SaaS teasers that speak by 0:15
4. Most first lines open on something other than the product
81 of the 111 first spoken lines in SaaS teasers open on something other than the product: the buyer's problem or pain (25), a bold claim (16), a scene, dialogue or instruction (15), a question (14), the customer or a persona (10) or a statistic (1). The other 30 lead with the product: 19 name the company or product and 11 introduce it. Every line was hand-coded against published definitions; the keyword lexicon written first agrees on 73 of 111, on all 14 questions, and on 12 of the 25 problem lines. The opening sentence is about the viewer's world; the product waits.
Companies whose first line opens on something other than the product include Papaya Global, Workable, Zapier, Lyzr, etc. The one statistic opener is Nooks, whose first line counts thousands of sales teams and billions in pipeline.
Epicor opens on the product's name at 0:00. Its February 2026 spot, 0:30 long, starts "Epicor Prism embedded into Profit 21 brings AI to your inventory team" over warehouse shots with AI with Epicor on screen. Where the product is an add-on to software the viewer already runs, the name is the news.
Source: ADVIDS Research coded set, 179 SaaS teasers from 150 companies, uploaded 6 May 2022 to 22 July 2026
Problem or pain leads by 6 teasers, and no single type holds a quarter of the set; the number is the opener almost nobody uses.
Beat 05 · Name · median at 0:07 · base: the 111 SaaS teasers that speak by 0:15
5. The product name arrives by 0:15, after the first line
77 of 111 speaking SaaS teasers say the product or company name by 0:15, at a median 0:07 among the 98 that say it at all, with the middle half between 0:04 and 0:14. Only 32 say it in the first line; in the 79 that do not, the name follows the start of that line by a median 0:10. 21 say it after 0:15 and 13 never say it in the transcript. On screen, the name or a logo card appears by 0:15 in 17 of the 39 visually coded teasers; of the 19 of those that also speak by 0:15, 6 name it both ways, 6 only aloud, 2 only on screen and 5 neither. The name is the second beat of the hook, not the first.
Companies that say the name aloud by 0:15 include The Trade Desk, Adobe, AssemblyAI, Emburse, etc.
Cohesity waits until 0:38 to say its name. Its May 2026 teaser, 0:57 long, opens "Identity attacks don't kick the door in. They log in." and plays out a lockout before a character says "We have Cohesity identity resilience". The problem runs for 0:38 of a 0:57 teaser before the name.
Source: ADVIDS Research coded set, 179 SaaS teasers from 150 companies, uploaded 6 May 2022 to 22 July 2026
Most names land between the first line and the end of the first 15 seconds, not on the first word.
Beat 06 · Words · median at 0:15 · base: the 111 SaaS teasers that speak by 0:15
6. Speaking teasers fit a median 33 words into 15 seconds
SaaS teasers that speak by 0:15 carry a median 33 words in the first 15 seconds, 2.2 words a second, with the middle half between 26 and 38. 73 of 111 speak 30 or more words, 19 speak 40 or more and 7 speak fewer than 20; the range runs from 8 to 54, the most from Paycom. Across all 179 teasers, counting music-only teasers as zero, the median is 24. On the 81 library videos with word-timed captions, the placed count by 0:15 is a median 1 word off the caption count. Fifteen seconds of a spoken SaaS teaser is a paragraph, not a slogan.
Companies whose teasers speak 30 or more words by 0:15 include Airia, Appian, Canopy, Swimlane, etc.
Corpay says 9 words in its first 15 seconds. Its March 2026 teaser, 0:30 long, speaks "Powered by AI, Corpay makes your business move" and then its name once more. The name sits inside the only sentence.
Source: ADVIDS Research coded set, 179 SaaS teasers from 150 companies, uploaded 6 May 2022 to 22 July 2026
Benchmarks
Descriptive figures for placing the first 15 seconds of your own SaaS teaser against the set.
The first 15 seconds, in medians
| Measure | Value | Base and range |
|---|---|---|
| Runtime, median | 0:41 | The middle half runs 0:30 to 0:52 |
| First spoken word, median | 0.3 s | Across the 111 teasers that speak by 0:15 |
| First on-screen text, median | 0:02.5 | Across the 39 visually coded teasers |
| Blank opening card, median length | 4.0 s | Across the 17 teasers that open on one |
| Name said aloud, median | 0:07 | Across the 98 speaking teasers that say it |
| Words spoken by 0:15, median | 33 | Middle half 26 to 38 words |
| On-screen words by 0:15, median | 13 | Across the 39 visually coded teasers |
| Coded segments by 0:15, median | 3 | Scene boundaries the coder marked in the first 15 seconds |
| Frame brightness, median | 117 | Mean luminance of the whole video, 0 to 255 |
What most teasers skip
| Pattern | Teasers | Base | Share |
|---|---|---|---|
| A statistic as the first spoken line | 1 | 111 | 1% |
| A logo card by 0:15 | 7 | 39 | 18% |
| A customer or persona as the first line | 10 | 111 | 9% |
| Introducing as the first line | 11 | 111 | 10% |
| The product interface by 0:15 | 19 | 39 | 49% |
The product interface sits just under half, so it is not mapped. Frame luminance splits three ways, 57 dark, 63 light and 59 mixed of 179, so no frame holds half; it is computed over the whole video, not the first 15 seconds. Companies with a logo card by 0:15 include Atlassian, ClayHog, Deliverect, Knak, etc.
What to do with this
- Beat 01 If 111 of 179 SaaS teasers speak by 0:15, decide in the brief whether yours is one of them, and write the captions for the other case.
- Beat 02 If the first on-screen text beats the voice in 26 of 39, design the first readable line as part of the opening, not as a caption added later.
- Beat 03 If 69 of 111 speaking teasers talk inside half a second, cut the music intro unless it carries the picture on its own.
- Beat 04 If 81 of 111 first lines open on something other than the product, choose deliberately between the problem, a claim, a question and the name, and sign off the first sentence on its own.
- Beat 05 If the name lands at a median 0:07 and only 32 of 111 say it in the first line, place the name inside the first 15 seconds and after the line that earns attention.
- Beat 06 If the median speaking teaser fits 33 words into 15 seconds, count the words in the first 15 seconds of your script before recording.
Methodology
| Vertical (visual record only) | Teasers |
|---|---|
| devtools | 6 |
| fintech | 3 |
| productivity | 3 |
| ai | 2 |
| cybersecurity | 2 |
| knowledge management | 2 |
| marketing | 2 |
| adtech | 1 |
| Other | 18 |
| Not recorded | 140 |
| Coding scheme | Values | Definition |
|---|---|---|
| Sound | voiceover · music only · sung or incidental vocal · voice after 0:15 | Audio kind comes from the library's audio analysis of the whole track. A spoken line is speech-segment text with two or more words that are not music or noise tags; lines that read as a sung track were coded by hand. |
| First word | seconds | The start of the earliest speech segment carrying a spoken line. Word-timed captions refine word times where a caption file aligns. |
| First line | problem or pain · product name · bold claim · scene, dialogue or instruction · question · introducing · customer or persona · statistic | The first sentence that starts before 0:15, cut at 25 words, hand-coded against written definitions with a fixed precedence. A keyword lexicon written first is kept as the first pass. |
| Name | seconds · by 0:15 · in the first line · never | The earliest placed word matching the company or product, including the spellings captions use for it, checked against each transcript. |
| Words by 0:15 | count | Transcript words placed before 15.0 seconds: caption word times where they align, otherwise each speech segment's words spread evenly across its span. |
| On screen | blank card or abstract ground · written type · people · interface · logo card · text by 0:15 · text before the voice | Coded per timed segment on the visually coded teasers: the opening segment's motif, whether it carries on-screen text, and the first segment carrying any text. |
| Frame | light · dark · mixed · brightness 0 to 255 | Computed, not coded: mean luminance over the sampled frames of the whole video, not the first 15 seconds and not the product's theme. |
| Not coded | cast · surface · altitude · build | The chapters rest on sound, speech, words and on-screen text; the library's build and surface codes cover only the visually coded teasers and are not used, and altitude is not recorded. |
Cite this report
Jai Ghosh, Advids. Decoding the SaaS Teaser: 179 Short-Form LinkedIn Ads Audited for the 15-Second Hook Strategy. 14 September 2026. https://advids.co/blog/saas-teaser-15-second-hook-audit
Author & editor bio
About Advids
Advids produces video for AI, SaaS and DeepTech companies. ADVIDS Research publishes coded studies of what those companies ship, with the dataset attached so every count can be checked.
Image credits and licensing
Video frames on this page are reproduced for analysis and commentary. Copyright in each frame remains with the company credited in its caption; to license a frame, contact that rights holder. Advids grants no rights over them.
Charts and images credited ADVIDS Research, and the coded dataset behind them, are © 2026 Advids and licensed CC BY 4.0: reuse them with attribution to Advids and a link to this page.