Video marketing 2026: Reels, VK Clips and YouTube Shorts - what to shoot and how to calculate the result
Comparison of video platforms for Russian business: coverage, CPM, formats. What to shoot in 20 minutes without a studio, how to calculate viewability and ERR. Niche benchmarks.
Most companies spend 80% of their video budget on production and 20% on distribution. This ratio is wrong - it should be the other way around. A studio video for 60,000 rubles with zero promotion will lose to a telephone video for 0 rubles, which was cleverly packaged in advertising.
I've been shooting videos for client projects on iPhone for three years now. Over the past 12 months, I have data on 7 projects where we compared studio content with telephone content: in 5 out of 7 telephone videos showed an ERR higher by 20-35%. Not because the studio content is bad - it’s just that the algorithm looks at retention, and telephone video is often perceived as more “own”, less advertising. I wrote this article using the figures of these projects plus data from colleagues in SMM chats.
“The first 3 seconds are not a “hook”, it’s either you kept your attention or you didn’t. The remaining 57 seconds are for those who remained."
1. Platforms and their audience in the Russian Federation 2026
The map of video platforms in Russia has become cemented over the past two years. YouTube has not gone away - 78 million unique users per month, of which 41 million watch Shorts. VK Clips grew to 52 million MAU and became the main tool for outreach video within the VKontakte ecosystem. Instagram with Reels has 38-42 million active users - despite everything, the audience has not left. Telegram video is a different story: there is no algorithmic distribution, the video only works for channel subscribers.
The difference in audience is not so much in size as in behavior. YouTube is a search engine, people come for an answer to a specific question. Shorts gets views from both search and recommendations. VK Clips are almost pure recommendations, like TikTok: a person is not looking for specific content, the algorithm itself decides what to show. Reels - mix: 60% recommendations, 40% subscriptions and search.
| Platform | MAU in Russia | Primary Audience | Average ERR | Advertising CPM | Screenability 60 sec |
|---|---|---|---|---|---|
| YouTube Shorts | 41 million | 18-45 years old, wide coverage | 1,5–3% | 150–400 ₽ | 28–38% |
| VK Clips | 52 million | 25-45 years, regions | 3–6% | 80–200 ₽ | 32–45% |
| Telegram video | ~30 million (views) | 25-40 years old, Moscow/St. Petersburg | 4–9% | no CPM format | 40–60% |
| Instagram Reels | 38-42 million | 18-34 years old, B2C | 2–5% | 200–500 ₽ | 25–35% |
Telegram is in a separate line here, because videos in TG are content for an already collected audience, not for cold outreach. The ERR there is high (4-9%) precisely because subscribers are watching - loyal people. I wrote more about building a channel from scratch in guide to Telegram channel 2026.
2. Formats: horizontal, vertical, square
Vertical format (9:16) is the standard for Reels, Clips and Shorts. You need to shoot vertically initially, do not reframe horizontally. The loss of quality when cropping is critical: the algorithm takes into account the percentage of full-screen viewing, and no one watches a cropped horizontal video in full screen.
Horizontal (16:9) lives on the YouTube channel in long format and in in-stream advertising. For YouTube Shorts, horizontal also works, but worse - the recommendation algorithm gives preference to vertical videos, all other things being equal. Square (1:1) is a dying format for Instagram feeds; it is almost never used for videos.
Duration: for Reels and Clips, the optimal viewability is 15-45 seconds, but the YouTube Shorts algorithm now better promotes videos from 45 to 90 seconds - they keep the user on the platform longer. My tests on three channels showed that 60-second Shorts collected 25-30% more impressions in recommendations compared to 25-second Shorts with the same ERR.
3. Production without a budget: filming and editing in 20 minutes
iPhone 13 and above, Samsung Galaxy S21 and above are enough. Light: shoot near a window during the day or buy a ring lamp for 1500-2500 RUB on the marketplace. Sound: the built-in microphone of a smartphone works at a distance of up to 80 cm. If you go further, a buttonhole for 600-800 rubles solves the problem. A tripod or flexible tripod for 500-800 rubles solves the problem of shaking.
Editing in 20 minutes is possible in CapCut: upload a video, trim off the excess, add subtitles (auto-generated, then edit in 3-4 minutes), add a track from the built-in library without copyright. Nothing else is needed - transitions, effects, animations only distract from the content. On four projects I tested “naked” videos against “dressed up” ones with a bunch of effects. The difference in ERR was in favor of simple ones - on average by 12%.
Subtitles are not an option, they are an obligation. 80% of users watch videos without sound in transport or in the office. A video without subtitles loses this audience completely. CapCut generates subtitles automatically in 30 seconds, then 3-4 minutes to correct errors - and that’s it.
4. Algorithms and what really affects coverage
The most common misconception is that the algorithm loves hashtags, posting time and regularity. All this works, but it’s secondary. One indicator is primary: audience retention. If people watch your video to the end, the algorithm distributes it further. If they leave in the first seconds, they don’t distribute it to anyone, no matter how many hashtags you put.
Specifically for signals that are publicly confirmed (Meta, YouTube) and tested in practice:
- The visibility of the first 3 seconds is the main signal on all platforms. The norm is 70%+, good from 80%.
- Average view duration / video length - for Reels the norm is 25-35%, good from 40%.
- Savings are the most expensive type of activity by weight in the algorithm, especially on Instagram.
- Reposts in Stories and personal messages - the Reels algorithm counts them and increases distribution.
- CTR on preview (for YouTube) - the optimal range is 4-8%, below 2% = a problem with the cover or title.
The time of publication has little effect for accounts with a small audience (up to 5,000 subscribers) - the algorithm itself will find the moment. For large accounts, reference point: Tuesday-Thursday 19:00-22:00 Moscow time. Hashtags: 3-5 thematic, no more - the algorithm perceives spamming hashtags as spam.
More details about construction SMM strategies for the Russian market taking into account the algorithmic priorities of each platform.
5. Metrics: viewability, ERR, transitions
Three metrics that I look at myself and that really tell me whether the video is working.
Screenability of the first 3 seconds = views lasting more than 3 seconds / total number of impressions × 100. This is a signal of how well the “hook” is working. Normal: 65-75%, good: 80%+. Below 50% - reshoot the first 3 seconds, the rest is not important.
ERR (Engagement Rate by Reach) = (likes + comments + reposts + saves) / reach × 100. Benchmarks by platform: VK Clips 3-6% norm, Instagram Reels 2-5%, YouTube Shorts 1-3%. Use everything above as advertising creative. Everything below is to analyze the audience and topic.
Link clicks / swipe up / click in bio — the final performance metric for business. There is no single rule here; it depends on the CTA in the video. My practice: if the video ends with a direct call (“link in the description”), a CTR of 0.3-0.8% of views is a normal result for niche B2B content. For B2C with a specific offer - 0.8-2%.
| Metrica | Weak | Norm | Okay | What does it affect? |
|---|---|---|---|---|
| Searchability 3 sec | less than 50% | 65–75% | 80%+ | does the algorithm distribute or not? |
| Average view share | less than 20% | 25–35% | 40%+ | coverage in recommendations |
| ERR (Reels) | less than 1% | 2–5% | 6%+ | organic distribution |
| ERR (VK Clips) | less than 2% | 3–6% | 7%+ | included in VK recommendations |
| CTR on the link | less than 0.2% | 0,3–0,8% | 1%+ | traffic to the site/landing page |
6. Video in paid advertising: in-stream and Reels Ads
Paid video advertising works in two scenarios: you promote an existing organic video or create a special advertising creative. The first way is cheaper - organic content with a high ERR in paid promotion receives a CPM that is 30-40% lower than a “cold” commercial. The algorithm sees that the content is already working with a real audience, and distributes it more readily.
YouTube in-stream (advertising before video) is one of the few formats where horizontal video retains its meaning. The first 5 seconds cannot be skipped, so the hook is needed there. CPM for in-stream: 150-400 ₽ depending on the niche, most competitive in finance and real estate (up to 600-800 ₽). View Rate (percentage of those who watched to the end or until the moment of skipping) the norm is 25-35%, good from 40%.
Reels Ads on Instagram - placement in the Reels feed, looks like a regular video marked “advertising”. CPM 200-500 ₽, CTR per landing 0.5-1.5%. Videos longer than 30 seconds lose here: the Meta algorithm optimizes for engagement indicators, and long videos are inferior in terms of watchability. VK Clips in the advertising account - CPM from 80 ₽, the cheapest video format on the Russian market right now.
Read about the variety of advertising formats on VKontakte in guide to VK advertising 2026. About creating video creatives for paid traffic - in the article about advertising creatives 2026.
7. Common mistakes
1. Start with a logo or greeting. “Hello, I’m Ivan, director of Romashka LLC, we are engaged in...” - this is a guaranteed drop in the first second. During this greeting, the user has already looked at three competitors. Start with the most interesting: with a result, with a question, with an unexpected fact.
2. Shoot a horizontal video and crop it vertically. Loss in quality, poor composition, faces are cut off. Shoot vertically initially if the goal is Reels, Clips or Shorts. If you need both a YouTube longread and Shorts, these are two different scenarios, two different filming formats.
3. Chase the number of platforms. “We need it everywhere” is dispersal without results. 12 good videos per month on one platform are better than 3 on four. The algorithm encourages regular authors, and it’s possible to pump up one channel—it’s almost impossible to pump up four at the same time with limited resources.
4. Don’t count views, look only at likes. Likes are a vanity metric. 10,000 views with 15% viewability is worse than 3,000 views with 55%. The algorithm sees everything, you only see likes. Open Analytics and look at the audience retention curve - you can clearly see at what second people leave.
5. Invest 80% of the budget in production. A studio video without promotion is money down the drain. Production = 20% of the budget, distribution = 80%. If you don’t have money for advertising, shoot cheaper and spend the extra money on promotion. A telephone video with 50,000 rubles per target will almost always beat a studio video without a distribution budget.
6. Publish one video per month and wait for the results. Short video algorithms work on volume. The minimum threshold for obtaining data is 12-15 videos. Only then does it become clear what topics and formats work for your audience. One video a month is a year-long experiment. During this time, competitors who publish 3 videos a week will go far ahead.
Conclusion
The video works - that's a fact. But not any video and not by itself. The formula is simple: cheap production + strong first 3 seconds + money for distribution + measuring the watchability of each video. Everything else is details. You can start today: shoot one vertical video on your phone, edit it in CapCut in 20 minutes, publish it with subtitles. After 48 hours - look at inspection. If below 50% at the 3 second mark, reshoot the beginning. That's what video marketing is all about.
Related materials: content plan template for SMM, SMM strategy for the Russian market 2026, VKontakte advertising 2026, short video script.
If you want to analyze your specific situation, write to Telegram or through form. Starting consultation - 0 ₽.