Как заработать на ВК шопс в 2026: охваты упали на 65%, что делать — Transcript
Full transcript
- 0:00In mid-October, an average VK shop was
- 0:02getting 120,000 views in three days. Two
- 0:05weeks later, it was down to just 12,000
- 0:08views. That is a 90%drop.
- 0:11The page I used as a benchmark is down
- 0:1383%.
- 0:15A regular post about jeans last week got
- 0:181,600 in three days without a single
- 0:21scandal. Something broke, but it didn't
- 0:23break for everyone. If you were just
- 0:26about to get into "shops," it probably
- 0:28feels like renting a kiosk by the subway
- 0:30, only for the station to close for
- 0:32repairs. Rent is due, the product is
- 0:33sitting there, but there are no people.
- 0:35In fact, the market is being purged, and
- 0:37only two paths remain. Many will be
- 0:39washed away now. Let me explain in more
- 0:41detail. The first path is a home
- 0:43shopping channel. Roughly speaking, a
- 0:45storefront page without any plot. Here’s
- 0:47the item, here’s the demo, here’s the
- 0:48link.
- 0:49And the second option is to raise a
- 0:51blogger, a thoughtful person whose life
- 0:53is followed and from whom people
- 0:55actually buy. Now I will show who the
- 0:58drop hit hardest and what is needed for
- 1:01each path.
- 1:02Doing both options manually is super
- 1:04difficult, so I’ll show you on screen
- 1:06how our AI agents do them.
- 1:09My name is Artem, and currently, 60 AI
- 1:11agents work for me. Some of them make
- 1:14shops, some write Zen articles, manage
- 1:16social media, and so on. You can’t
- 1:19compare a fresh post with an old one.
- 1:20The old one simply hung around longer.
- 1:22Therefore, I am taking a consistent age
- 1:24: three days after publication. Posts
- 1:27from October 12th—12,500 median, from
- 1:30the 19th—6.2 k, from the 26th—4,300.
- 1:35The whales: 100,000 in three days used
- 1:37to be reached by every seventh shop, now
- 1:39by every fifteenth.
- 1:40You’ll probably say, "There’s just more
- 1:43content now." I counted 579 shops in the
- 1:46first week. 582 in the last; in fact,
- 1:49it’s the same amount. And this is where
- 1:52it gets interesting.
- 1:54Since October, I’ve been teaching our AI
- 1:56agents one thing: direct advertising
- 1:58doesn’t sell, stories do.
- 2:01The models were a few channels: a
- 2:03scandal in the first line, a husband, a
- 2:05mother-in-law, the product in the middle
- 2:07, and arguments in the comments.
- 2:09Here is an example of such good channels
- 2:11. A: 260,000 in three days, now 45k,
- 2:15down 83%.
- 2:17Second channel 372,000, now 7.5 k, down
- 2:2198%.
- 2:24And now the most boring pages, with
- 2:26titles like "finds,""reviews,""unboxing
- 2:28,""discounts."
- 2:29Grouping by title is crude, but the
- 2:31difference is visible.
- 2:33They lost 18%, all others 15%.
- 2:38And storefronts outperformed the rest.
- 2:406,000 in median versus 5,900.
- 2:44Ozon finds actually grew threefold.
- 2:47The most boring format survived the drop
- 2:49the best. So that’s where we’ll start.
- 2:51Remember home shopping TV? The host, a
- 2:53frying pan. Call right now. But that’s
- 2:56not all. Everyone laughed. But they
- 2:58bought the frying pans because there’s
- 3:00nothing there except a demonstration.
- 3:02Here’s the problem, here’s the item,
- 3:04here’s how it works. Last week, two
- 3:06ordinary posts really took off. A woman
- 3:08writes that she stopped seeing the marks
- 3:11on her outdoor thermometer and shows an
- 3:13electronic one. 622,000 views in three
- 3:16days. Another one just shows some jeans
- 3:19. 600,000. No mother-in-law, no drama.
- 3:22What do you need for this? A product
- 3:24costing from 200 rubles. up to 1,000.
- 3:26The commission there is from 15%.
- 3:28Stock of more than 1,000 units. On the "
- 3:30Check on Shops" podcast, a guest talked
- 3:33about a leaf blower. 10 million views
- 3:35and five sales. The most common failure
- 3:38story. Or one goes viral, but the
- 3:40product sells out in hours. Three shots
- 3:43. The first is lively and a bit quirky.
- 3:45The second is the item in action. And
- 3:47the third is a close-up of the item.
- 3:48Regarding text: pain, solution, what
- 3:51happened after. 700-900 characters and
- 3:53no words that moderation dislikes. And
- 3:57also pace. Virality is unpredictable.
- 3:59Some shops break even, even for strong
- 4:02authors. Chart leaders post 90 pieces a
- 4:05month, or even more. I've compiled
- 4:07product criteria, a post template, and a
- 4:09list of words into one file. Grab it for
- 4:11free. The first link in the description
- 4:13looks like the format has been found.
- 4:16Take the couch advertising. But that's
- 4:18not all. A storefront has no one but the
- 4:21algorithm. Today the feed shows it,
- 4:23tomorrow it stops. And there isn't a
- 4:25single person who will visit the page on
- 4:27their own. But look who has grown among
- 4:29the big pages.
- 4:31These are all pages with real people or
- 4:34characters.
- 4:36In fact, reach has dropped for everyone
- 4:38, but those with a face have a much
- 4:41higher potential for people to stay and
- 4:43buy. But a face means filming every day
- 4:47for years, or the face can be made up.
- 4:49And a blogger isn't just a picture of a
- 4:51pretty girl. It's a person who was
- 4:53created entirely from scratch.
- 4:55I have a grandma, she's 61, retired,
- 4:57works at a parcel pickup point, and
- 5:00she's currently in a whole season
- 5:02regarding a 186,000 credit card debt.
- 5:06What is needed? A biography with some
- 5:08mystery for those who read and buy. And
- 5:11that's age 45+. Adult children, pension
- 5:14, summer house, one face in all photos.
- 5:17I stumbled on this myself. In every shot
- 5:19, the character was slightly different,
- 5:21and it's immediately visible. Every day
- 5:23a post without a product, about family,
- 5:25money, voting. The shop fits into this
- 5:28series as a part of life. I think the
- 5:29two paths are clear. Now let's calculate
- 5:32what this costs you. One shop with
- 5:34hands-on work.
- 5:35Opened the catalog, sorted through cards
- 5:37, checked the price, commissions, stock
- 5:39, took three shots, wrote the text,
- 5:40uploaded the application, wait up to a
- 5:42day for moderation, and get a rejection
- 5:44.
- 5:44I had one product rejected twice this
- 5:46way. First for a provocative headline,
- 5:48then, apparently, for promising results
- 5:51and before-and-after photos. Plus, you
- 5:54can only keep three unscheduled requests
- 5:55on the page. And so it goes in circles.
- 5:58Even if a shop takes just an hour, three
- 6:00a day is already 3 hours after work.
- 6:0390 hours a month with no days off. And
- 6:06most of these shops sell little or
- 6:09nothing. This is the point where you
- 6:11want to quit, because in the evening
- 6:13there’s TV, chores, and so on.
- 6:15It was exactly the same for me until my
- 6:17agent took it over. Look, a product just
- 6:19arrived in my Telegram. The agent found
- 6:22it in the cabinet recommendations itself
- 6:24, checked the price, commission, and
- 6:25stock.
- 6:26I effectively have two buttons: make a
- 6:28Shoops or choose another product. And
- 6:30here is a finished Shoops. Three photos
- 6:32and text. The face is the same. Stop
- 6:34words. None. It learned moderation rules
- 6:36from hundreds of real requests. If I
- 6:38don’t like a line, I reply with text. In
- 6:41plain words, like with an employee. I
- 6:43like it. Second button. Send for
- 6:45moderation. It fills out the request
- 6:47itself. And here it is working in the
- 6:49cabinet right now. I’m not touching
- 6:52anything. Rejected. It reads the reason
- 6:55, fixes it, and sends it again, just
- 6:57letting me know what it changed.
- 7:00Approved, it sets the time itself; when
- 7:01it’s up, it removes the label and
- 7:03searches for the next product. It checks
- 7:05everything every half hour from 8:00 AM
- 7:08to 3:00 AM. Installation is just: "
- 7:10Unpack the folder."
- 7:12Open the app or code, write the settings
- 7:14.
- 7:15It checks the program itself, creates a
- 7:17community character, and offers options
- 7:19for you to choose. Every evening at 7:00
- 7:21, it posts to keep the page active. No
- 7:24new Shoops will go out without your
- 7:26button press.
- 7:27Those 90 hours turn into just a couple
- 7:30of taps for each shop.
- 7:31All of this is fully automated. Product
- 7:34, text, photos, requests. And after
- 7:36moderation, the agent does everything.
- 7:38You just press buttons, and you’ll ask:
- 7:41"How much will this bring in?" I won’t
- 7:43promise you a specific sum. The agent
- 7:45takes the work and routine off your
- 7:46shoulders, but it doesn't guarantee
- 7:48orders. Here are two figures from the
- 7:50same podcast. Valeriya Sokolova made her
- 7:52first million with 2,000 subscribers.
- 7:54Oksana Morozova had one post with 20,000
- 7:57reach that brought in 250,000 rubles.
- 8:0020,000 views is not actually that many.
- 8:03Meaning, it’s not reach that feeds you,
- 8:05but the item that sells. And you can
- 8:07only find that through the number of
- 8:09attempts. By hand, that’s, say, 10
- 8:11attempts a month, if you don't quit.
- 8:13With the agent, it’s up to 90 per page.
- 8:17The agent is in my "Money Agent" club.
- 8:20Inside is the VK Shoops agent,
- 8:21instructions, mini-training on shops for
- 8:24beginners, moderation rules, lessons on
- 8:26neural networks, and a Zen agent as a
- 8:29gift.
- 8:30What will you have to do? The computer
- 8:32needs to be turned on for this agent.
- 8:34Subscriptions for AI and images, if you
- 8:37want the paid ones, are about 15-20
- 8:40dollars. If you're expecting money next
- 8:43week, this isn't for you. Scroll on.
- 8:46If you're ready to click buttons for a
- 8:47month and see what items sell, the link
- 8:48is the first line in the description.
- 8:51There’s a free file there too.
About this transcript
This page contains the full transcript of Как заработать на ВК шопс в 2026: охваты упали на 65%, что делать by Артём Дубровец | ОТКУДА БАБКИ?, generated from the public captions YouTube serves with the video. The transcript has 1,529 words across 235 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.