Katarza
Katarza is a regional subscription platform for short business videos from practitioners, what its founder calls a "business Netflix": short clips and practical answers from people who have already solved the problem. The most important thing we built for it is a social media system. It turns the platform's own videos into stories, carousels, static posts and reels for Instagram, TikTok, LinkedIn and YouTube, schedules them by written rules, reads the results of everything already published and moves the next posts to where they work best. People approve what goes out. Around that system sits the rest of our work since launch in December 2025: consulting, positioning, measurement, email and a separate offer for companies.
A subscription platform depends on two things: whether new people find a reason to start, and whether subscribers find a reason to stay. Katarza had the library for both. Practitioners talk about taxes, contracts, teams and clients, filmed by a team of film producers, so it looks and feels like film. The free MikroKAT and MiniKAT formats show what is inside, and MasterKAT is the paid format that goes deep.
By late September 2026 the hardest question was on the table: individual subscriptions were not coming in the way the founder expected. Our diagnosis was that video made for the platform is not native to social media. It has no fast cuts, no trigger and no hook in the first seconds. Posts were being put up by hand from a shared spreadsheet calendar, and every practitioner brings a different audience, voice and topic. The founder gave us a free hand on social, and from mid-October we took over the calendar and the content for every channel and replaced the spreadsheet with a system.
The source is always the platform itself. The speech in finished reels from Katarza's Drive is transcribed, so every caption and every quote rests on what the practitioner actually said. From there a set of small builders produces the formats, by the same rules for every video:
The calendar runs on written rules, checked before every push and after every move: a daily limit per channel, fixed slots per format, one practitioner at most once a day on a channel, never a reel and a carousel of the same person on the same day, and Tuesday reserved for whatever is new on the platform. Captions on Instagram and TikTok are different because the audiences are different, and there are no hashtags. Whatever does not fit moves to the next month. Nothing is deleted to make room.
Once a week the system pulls every published post and its metrics from Buffer and compares like with like: same channel, same format, broken down by person, hour, day, opening line and whether the practitioner was tagged. It works with medians and ignores any group of fewer than four posts.
The first full pass, in November 2026, changed several things at once. Stories did better at midday than in the evening, and tagging the practitioner turned out to be the strongest lever. Almost everyone who sees the first frame of a story stays to the last one, so the real question was how many people see the first frame at all. On TikTok and LinkedIn a statement worked better than a question as the opening line.
It also showed where we got it wrong. TikTok views dropped in the exact week posts moved from late evening to midday, and hand-made and system-made posts dropped equally, so the cause was the hour. TikTok moved to the early evening, and the effect gets checked at the start of December. On the Instagram feed, posts built from the first templates did worse than hand-made ones, and the diagnosis points at the template. And once a post moved by hand in Buffer left its old slot marked as taken, so the next push skipped two reels. Since then a post moves only through one command that updates Buffer and our schedule in the same step.
On the first of every month a routine runs the same analysis without anyone starting it, saves a snapshot and compares it with the month before. A signal becomes a rule only when it holds two months in a row, on at least eight posts, with at least one and a half times the median for that channel and format. A signal seen in one month only goes into the report as something to watch. Before accepting a rule, the routine checks whether the strong group comes from a single month or a single person. When a channel and format falls 20 percent or more month over month, it looks for the cause (hour, template, volume, ads, tagging) and never stops at "a general dip". Then it moves already scheduled posts out of weak slots into strong ones, under the same calendar rules, with a dry run first, and reports what it moved.
The system decides where a post goes, never whether it goes. Nothing is scheduled until someone on our side runs the push with an explicit go, and practitioners are tagged by hand, in posts and in stories. The routine may move posts between slots because a move can be undone. It may not delete a post, change the text or media of a scheduled post, add new posts or change how many go out per day; those come back as proposals with the exact command, waiting for a yes. Approving content and anything that cannot be undone stay with people.
The social system sits on top of the commercial work we have done with Katarza since launch. The first piece was measurement. Ads were sending people to generic pages and the data showed no purchases at all. When we went through the tag setup, it turned out that both purchase tags were dead, so nobody could see what was working. We rebuilt purchase tracking in GA4 and added purchase and registration events to the Meta pixel, so campaigns could finally see a subscription.
The second was the page and the size of the ask. We built a sales page that starts from a problem and leads to the practitioner who has solved it, and we split the path in two. People who had never heard of Katarza were invited to a free webinar first, and only warm audiences went to the subscription page. Over time the pattern in the ads was clear: video with a face and a story brought subscribers, while faceless visuals and carousels brought none.
The third was positioning. The founder told us that the way Katarza had been described did not reach the audience. From September it presents itself as a practical tool for business and career problems: short videos, answers and direct contact with practitioners. Every piece of copy now opens with a concrete problem, such as a letter from the tax office, a contract or a first hire, and only then gets to what is on the platform.
Alongside that we built a separate page for companies that want to give their teams access, a code program that gives a free first month through each practitioner's recommendation, and an email plan measured by replies and clicks. Opens do not count, because Apple Mail inflates them, and email is the only place where Katarza talks to its subscribers directly.
Katarza now has a social media system that turns the platform's own videos into posts for four channels, schedules them by written rules, reads the results every week and every month moves the slots based on what earlier posts did, without anyone having to start it. People approve what goes out and anything that cannot be undone. Around it sits the rest of the commercial system: measurement that sees a subscription, a free path into the paid format, a separate offer for companies and email measured by replies. We do not publish the platform's numbers, so the result we stand behind is the system and the way it corrects itself.
Who we are
Strategy, marketing, sales, customers and the technology under them, under one roof and with the same people from day one.
Who we build for