Open a mixed-media tech demo and you usually see a picture first. A factory floor. A lab bench. A chip, rendered in blue. The usual brief says that is what premium looks like. Picture wall to wall, product saved for the end. But in 57% of videos, the product screens get more time than the picture. So how many seconds does your brief give the screen your buyer will use?
A mixed-media tech demo mixes a picture with other media. The picture means camera shots, stock shots or 3D renders (CGI). The other media are the product screen, motion graphics or 2D animation, and moving text on screen. ADVIDS Research coded 74 of these tech demo videos frame by frame, from OSHIMA, Zoho, HPE, Illumina and others. Whether a mix looks Hollywood-grade stays your call. The files hold no budget or quality data.
Seven patterns stand out. A product screen joins the picture (Paired Screen). Live action shows up more than 3D (Camera First). The picture opens and the screen follows (Scene Handoff). Screens fill the whole frame (Full Frame). Screens get more time than the picture (Screen Time). A voice explains it (Voice Track). And most frames are dark, often blue (Shared Ground).
ADVIDS Research · 74 B2B tech demo videos · 62 companies · coded frame by frame · published 14 September 2026 · 20 min read
Top takeaways
Paired Screen
82% of the B2B tech demo videos pair the picture with a product screen.
Camera First
62% use live action or stock shots, and 35% use 3D CGI.
Scene Handoff
69% of videos with both a picture and a product screen show the picture first.
Full Frame
79% of videos with a product screen let it fill the whole frame.
Screen Time
57% give the product screens more time than the picture.
Voice Track
73% use a voiceover, and 11 of the 20 with music only use moving text instead.
Shared Ground
73% run on dark or mixed frames, and blue is the top main color.
Questions this report answers
What is mixed media animation in a B2B tech demo video?
It mixes a picture, meaning live action, stock or CGI, with other media: a product screen, motion graphics or 2D animation, or moving text. ADVIDS Research coded 74 of these videos from 62 companies. 78% use three or more kinds of media. That's the Paired Screen: a product screen joins the picture in most of them.
Do B2B tech demo videos use live action or CGI?
Live action more often. That's Camera First. 62% of the 74 videos use live action or stock shots, and 35% use 3D CGI. A few use both. CGI alone fits a product no camera can reach, like the inside of a chip. Intel's three Foundry videos are all rendered.
What order do cinematic B2B demo videos run in?
Picture first, then product. That's the Scene Handoff. Of the videos with both, 69% show the picture before the product screen. The first screen typically lands 16 seconds after the picture starts. A title or logo card closes most of them.
How does the product interface appear in a cinematic tech video?
Mostly full frame. That's the Full Frame pattern. Of the videos with a product screen, 79% let it fill the whole frame. Some also show it on a laptop in the shot, like Zoho's Cliq video. Give the whole frame to any product screen your buyer has to read.
How much of a mixed-media product video should show the product screen?
More than the picture, if you sell software. That's Screen Time. In 57% of the videos, screens get more time than the picture. In a typical video, screens fill 33 percent of the time and the picture 24 percent. Machines are the exception. There, the hardware gets most of the time.
Do cinematic B2B product videos use a voiceover?
Yes, mostly. That's the Voice Track. 73% of the videos use a voiceover. The other 20 run on music, Crestron and Symantec among them. 11 of those put their words on screen as moving on-screen text. Music alone works only when every claim fits one line of text.
Should a mixed-media tech demo video be dark or light?
Pick one look for the picture and the screens. That's Shared Ground. 73% of the videos use dark or mixed frames, and blue is the most common main color. If your product screen is light, keep the picture light too. SAP's AI Agent Hub overview does this.
How was this analysis done, and what does Hollywood-grade mean here?
ADVIDS Research checked 188 videos against a rule written before the work began. 74 qualified, from 62 companies, 6 short of the usual 80. Each was coded frame by frame. Seven patterns stand out: Paired Screen, Camera First, Scene Handoff, Full Frame, Screen Time, Voice Track and Shared Ground. Hollywood-grade is your test. Nothing here rates quality.
How to read this report
This report looks at 74 product, demo and explainer videos from 62 tech companies, uploaded between 14 January 2026 and 3 August 2026. Each figure counts videos, like 61 of 74. That is 6 short of the 80 this series usually wants, so read the figures as a guide.
Pattern 01 · Paired Screen
1. Most videos pair the picture with a product screen.
Most mixed-media tech demos don't choose between a picture and the product. They show both. 82% put a product screen next to the picture (61 of 74). And 78% use three or more kinds of media. So the question isn't whether to mix. It's which media. If your buyer uses your product on a screen, show that screen.
Think of a hotel brochure. The pool sells the stay. But the guest lives in the room. Your product screen is the room.
The most common mix uses all four: picture, screens, motion graphics or 2D animation, and moving text. 29 videos do that.
Example: OSHIMA's video for its TAC cutting machine uses all four. At 0:20, a monitor shows cut patterns in front of the machine. Machine and screen, in one shot.
Picture plus a product screen: companies in this group include SAP, Crestron, Bloomfire, Cadence, etc.
Exception: hardware sold at a counter. Global Payments' Genius for restaurants video has no product screen at all. Restaurant shots, the device and moving text carry it.
Source: Advids analysis of 74 B2B tech demo videos
Each video sits in one mix. The seven add up to 74.
Pattern 02 · Camera First
2. Live action shows up more often than 3D CGI.
If a camera can see the product, most of these videos use one. 62% use live action or stock shots (46 of 74). Only 35% use 3D CGI. So if your product has a body a lens can see, put it in front of a camera first. Build it in 3D only when no camera can get there.
A camera records a room that already exists. A render has to build the room first. Your robot arm already exists.
Most use the camera alone: 38 videos have no CGI at all. Camera b-roll or stock and product or hardware shots make up most of the picture.
Example: Lenze's motion-control video is on camera from 0:04, with music and no voiceover. At 1:15, a robot head picks products off a blue-lit line. Real arm. Real line.
Live action or stock: companies in this group include AFL, Amplitude, Cavli Wireless, Emburse, etc.
Exception: a product inside a chip. Intel's Omni-MIM video runs on 3D renders of chip layers. There is nothing there for a camera to see.
Source: Advids analysis of 74 B2B tech demo videos
The four groups add to 74. The 10 with neither still show a picture through other shots.
Pattern 03 · Scene Handoff
3. The picture opens, then the product screen follows.
Most videos set the scene before they show the product. Of the videos with both a picture and a product screen, 69% show the picture first (35 of 51). The screen follows soon after. If your buyer needs a scene before the product makes sense, open on the picture. Then bring the screen in early.
A play opens on the set before anyone speaks. Your picture is the set. Your first product screen is the first line.
The first screen typically lands 16 seconds after the picture starts. And a title or logo card closes most videos.
Example: Jimi IoT's tracker video opens on the product. At 0:10, the rendered tracker sits on a dark gray ground. The app arrives on a phone 21.5 seconds in. Scene, then screen.
Picture first: companies in this group include IBM, Epicor, Genpact, HCLTech, etc.
Exception: the product is the canvas. Celtra's AI Creative Flows video opens on its node canvas at 0:02. When the canvas is the product, open on the canvas.
Source: Advids analysis of 74 B2B tech demo videos
The four add to 74. Title cards and presenters are not counted here.
Pattern 04 · Full Frame
4. Most product screens fill the whole frame.
When the product screen appears, it usually fills the frame. 79% of the videos with a screen show it full frame (48 of 61). Some also place it on a laptop or phone inside a shot. That shows where the product lives. But if your buyer has to read the screen, give it the whole frame.
A menu across the room tells you the place serves food. The menu in your hand tells you what to order.
Screens come two ways: an animated interface mockup or captured product screens. Mockups show up more often, in 28 videos against 18.
Example: Zoho's Cliq video opens inside the shot. At 0:03, the camera looks over a shoulder at a laptop with Cliq open. The screen isn't there to be read. It puts the product on a real desk.
Full frame: companies in this group include HPE, AIMS360, Amplience, Arm, etc.
Exception: every screen has fields to read. Rillion's Capture video cuts between full-frame invoice screens and a woman at a laptop. The screen never enters her shot.
Source: Advids analysis of 74 B2B tech demo videos
The three add to 61. Inside the shot means on a device in the shot, or laid over the picture.
Pattern 05 · Screen Time
5. Product screens get more time than the picture.
The picture may open the video, but the screens hold it. In 57% of the videos, product screens get more time than the picture (42 of 74). If you sell software, give the screens the most time. The picture sets the scene. The screens show what your buyer is paying for.
Think of a home listing. One photo of the street, then every room. Your picture is the street.
In a typical video, screens fill 33 percent of the time and the picture 24 percent. Moving text and motion graphics get less.
Example: HPE's Data Fabric video opens on an office. At 0:42, a resource page fills the frame with capacity tiles. HPE gives its dashboards 60 percent of the time, and the picture 32.
Screens ahead of the picture: companies in this group include Contentstack, Google Cloud, Infobip, Jodoo, etc.
Exception: the product is a machine. ESSERT Robotics' cleanroom video gives robot hardware 87 percent of its time. The machine is what you buy, so it gets the time.
Source: Advids analysis of 74 B2B tech demo videos
Each bar is the typical share in videos that use that medium, so the bars don't add to 100.
Pattern 06 · Voice Track
6. Most videos use a voiceover to explain.
A picture rarely says what a product does on its own. So most videos add a voice. 73% use a voiceover (54 of 74). The rest run on music and often put their words on screen instead. If a claim won't fit one short line of text, record a voice for it.
Think of a museum audio guide. The painting hangs on the wall. The voice tells you what you're looking at.
Of the 20 that run on music with no voiceover, 11 use moving text. A few have captions burned into the picture instead.
Example: Illumina's Insight Ecosystem video opens on lab shots. At 0:16, a gloved hand places a sample at an instrument, with no words on screen. The voice does the explaining.
Voiceover: companies in this group include Doxis, Global Payments, Rackspace Technology, Samsung, etc.
Exception: the idea is a picture. Symantec's CBX video runs on music only. 3D amplifiers and gauges carry the labels Protection and Uptime, so the labels do the talking.
Source: Advids analysis of 74 B2B tech demo videos
The three add to 74.
Pattern 07 · Shared Ground
7. Most videos run dark, and blue is the top color.
Most of these videos look dark. 73% use dark or mixed frames (54 of 74). And blue is the most common main color. If your picture and your screens cut back and forth, pick one look for both. A jump from dark to light can make one video feel like two.
A play doesn't jump from noon to midnight in the same scene. Your picture and your screens are the same scene.
Blue is the main color in 29 videos. Warm tones and green trail far behind. And 20 videos run light.
Example: AFL's splicer teaser sits at the dark end. At 0:03, the Fujikura splicer glows deep blue on a black ground. Its product screen arrives on a handheld device at 0:12.
Dark or mixed frames: companies in this group include Showpad, SSI Schaefer, SuperOffice, Symantec, etc.
Exception: a product with a light screen. SAP's AI Agent Hub overview runs on light frames, around light chat panels and dashboards. The picture stays light to match the product.
Source: Advids analysis of 74 B2B tech demo videos
The five add to 74. Grays and blacks don't count as a main color.
Benchmarks
Check your own video against each row. Each one is the typical value across the 74.
Length and timing
| Measure | Typical (74 videos) | What it means |
|---|---|---|
| Runtime | 1:14 | Half the videos run shorter than this |
| First picture shot | 0:07 (64 videos) | When the picture first appears, in videos that have one |
| First product screen | 0:17 (61 videos) | When the product screen first appears, in videos that have one |
| First moving text | 0:07 (34 videos) | When moving text first appears, in videos that have it |
| First logo card | 1:00 (49 videos) | When the brand logo first appears, in videos that have one |
| Words spoken | 104 | Words in the voiceover; music-only videos count as zero |
In a typical video, the picture and moving text arrive at 0:07, and the product screen at 0:17.
What almost nobody does
| Pattern | Videos | Base | Companies |
|---|---|---|---|
| Product screen laid over the picture | 13 | 74 | SAP, Cavli Wireless, Cegid, Contentstack, etc. |
| Opens on the product screen | 13 | 74 | G-Stacker, Rillion, RMS, Sana, etc. |
| Presenter on camera | 10 | 74 | Bloomfire, Emburse, G-Stacker, Jodoo, etc. |
| Product screens fill half the time or more | 10 | 74 | Bloomfire, HPE, Sana, SAP, etc. |
| Captions burned into the picture | 5 | 74 | Cegid, G-Stacker, Ofelia, OSHIMA, etc. |
Each of these shows up in fewer than a quarter of the videos.
Seven questions for your next brief
Answer these seven in writing before anyone books a camera or a render.
- Pattern 01 Will your buyer use the product on a screen? Brief the screens and the picture of your B2B tech demos as one job.
- Pattern 02 Can a camera see your product? Put it in front of a lens before pricing renders.
- Pattern 03 Is your product a scene or a canvas? Scene: write the second the first screen lands into the video script. Canvas: open on it.
- Pattern 04 Does your buyer have to read this screen? Give it the full frame, and plan those screens before production is scheduled.
- Pattern 05 Do you sell software or a machine? Software: the screens get more time. Machine: the hardware does.
- Pattern 06 Is a voice booked? Record it before the final edit. Music only: every claim gets its own line on screen.
- Pattern 07 Is your product light or dark? Pick one look for the picture and the screens in the creative brief.
Now go back to that opening picture: the factory floor, the lab bench, the rendered chip. It opens most of these videos. But in 42 of 74, the screens hold more of the time.
This week, mark every screen in your script read or recognize. Then time where the first one lands.
The picture opens the video. The screen holds it.
Methodology
| Vertical | Videos |
|---|---|
| retail | 4 |
| fintech | 4 |
| semiconductors | 4 |
| enterprise | 4 |
| cx | 3 |
| logistics | 2 |
| hospitality | 2 |
| other verticals | 51 |
| Coding scheme | Values | Definition |
|---|---|---|
| Cinematic picture | cinematic format · live_action · stock · 3d_cgi · cinematic_broll · office and crowd stock · product_shot · robot_hardware · cgi_environment · facility_render · manufacturing | Coded from frames: camera or rendered picture. Live action and stock count only when the format is not talking_head. A product_shot segment can be a render or a mockup the coder framed as a product. |
| Interface | ui_mockup · screen_recording · screencast · dashboard_ui · chat_ui · terminal_ui · handheld_device_ui · notification_alert · screen_recording_generic | Coded from frames: the product's screens, animated as a mockup or captured from the product. |
| Motion graphics or 2D | motion_graphics · 2d_animation · diagram_network · diagram_flowchart · data_chart · illustration motifs | Coded from frames: drawn graphics, diagrams, charts and illustration. |
| Kinetic type | kinetic_type as format, build code or segment | Coded from frames: animated words as a medium. Title cards and logo cards do not count. |
| Interface entry | device · overlay · full frame | Read from the frame-coded label of each interface segment: device when the segment is a handheld-device motif or its label names a laptop, phone, tablet, monitor, panel or over-shoulder view; overlay when the label names an overlay, hologram or floating panel; full frame otherwise. |
| Order | first medium · first cinematic and interface timecodes · closing segment | Computed over timed segments. The first medium is the first segment that is not a card, presenter or abstract background. |
| Runtime share | percent of runtime | Computed: the union of a family's timed segments over runtime. A family present only in format or build codes has no segment and no share. |
| Delivery | voiceover · music only · presenter · captions burned in · word count | Voiceover or music only comes from audio analysis; presenters and burned-in captions are coded from frames; word counts come from auto or manual captions. |
| Frame | light · dark · mixed · brightness 0 to 255 · lead hue | Computed, not coded: mean luminance and palette across the nine sampled frames. It covers every frame, not the product interface alone. |
| Not coded | cast · surface · altitude | Who appears, how much of the product surface shows, and what each video argues. The library records none of them exactly, and this report does not code them from transcripts. |
Cite this report
Jai Ghosh, Advids. The Cinematic Report: We Analyzed 74 B2B Tech Demos for Hollywood-Grade Mixed Media Animation. 14 September 2026. https://advids.co/blog/cinematic-mixed-media-b2b-tech-demos
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.