In short
- A free iOS market-share chart became a press magnet nobody else was giving away. TelemetryDeck's public iOS version-adoption survey pages turned into a go-to source cited by 9to5Mac, Forbes, iOS Dev Weekly, and the Wall Street Journal — free data doing the work most companies pay PR agencies for.
- It took 18 months of unpaid groundwork before the first paying customer showed up. Founded when Covid canceled a world trip in spring 2020, TelemetryDeck didn't land a paying customer or its first outside funding until October 2021.
- Usage exploded 146,000% in one year, but revenue took until 2025 to clear €100K. 577 million signals were processed in 2022 alone — yet the company narrowly missed €100,000 in annual sales for 2024, and only called itself 'just about profitable' in May 2026.
- By 2026, AI recommendations quietly replaced social media as the top acquisition channel. Claude Code and ChatGPT became the most important systems recommending TelemetryDeck, after years of social-media and forum presence — the same public trail, read by a new kind of visitor.
The mechanism
TelemetryDeck's growth chain runs: publish a genuinely useful public data asset (the iOS market-share survey) that journalists have real incentive to cite for free, because nobody else offers it; stay genuinely present in the specific developer-native communities (Reddit, Mastodon) where that press exposure can convert into trust rather than evaporating; and because that public trail of survey pages, press citations, and forum presence compounded for years before AI search existed, it became exactly the material LLMs now read by default — so the 2026 shift to AI-driven recommendations is a new reader finding an old trail, not a new mechanism replacing the old one.
Giving away a genuinely useful public data asset made TelemetryDeck a source journalists cited for free
Starting around November 2022, TelemetryDeck began publishing iOS version market-share charts as blog posts, then moved them to dedicated landing pages. This wasn't a marketing stunt — it was a byproduct of data the company already collected — but it filled a real gap: nobody else was publishing this specific, continuously updated dataset for free. By 2026, major outlets were treating it as a default citation, which gave TelemetryDeck recurring, unpaid press exposure that a small team could never have bought.
We are super proud that 9to5Mac, Forbes Magazine, iOS Dev Weekly, and the Wall Street Journal are using TelemetryDeck as a top source for insights about iOS market share. We are also a recommended source on several newsletters, on Reddit, and on a lot of other technology websites.
Showing up genuinely in developer-native communities (Reddit, Mastodon) built the trust that made the press citations convert into signups
A press mention alone doesn't create a customer — someone still has to trust the tool enough to install an SDK in their app. TelemetryDeck's founders tracked their channels directly and found social media was their single most important one overall, with Mastodon sending more visitors than any other social network, while a custom crawler (Redreach) kept them present in Reddit conversations about app analytics without paid promotion. That sustained, unpaid presence is what turned occasional press citations into a steady stream of developers who already recognized the name.
Social media is not our only channel, but it is our most important one. Looking at social networks alone, Mastodon sends us the most website visitors. Reddit, though, has been our secret weapon. We keep an eye on mentions of TelemetryDeck or conversations about app analytics through Redreach. The result? Steady traffic (Reddit is always in the top 10 referrers for our website).
The same public trail that fed journalists and forums for years is now what AI systems read by default
By early 2026, the founders noticed that new-customer recommendations had stopped coming primarily from social media and were instead coming from LLMs. This wasn't a new content strategy replacing the old one — it was the years of survey pages, forum mentions, and press citations becoming exactly the kind of public, citable material that AI systems like Claude Code and ChatGPT default to when asked about app analytics tools. The founders actively adapted their website from August 2025 onward to make this trail easier for AI to read, but the underlying asset it was reading from was the same one built for links 1 and 2.
Since the beginning of 2026, this has completely changed. Recommendations are still important, but they are now given by LLMs instead of from social media platforms. Claude Code and ChatGPT are the most important AI systems where we are recommended or found.
How it went
Founding through the first paying customer
2020-04 → 2021-10TelemetryDeck began when Covid canceled Daniel Jilg's world trip, giving him time to finally build the lean, privacy-first analytics tool he had always wanted as an app developer [S1][S2][S12]. Lisa Figas, who had been Daniel's friend since they met on Twitter in 2012, joined as co-founder [S5]. Over the next 18 months a small team — 2 founders plus several part-time helpers covering the Viewer app, Android and Unity SDKs, and the web dashboard — registered the company, rebranded from an earlier name to 'TelemetryDeck' in September 2021, and applied for a Bavarian founders' prize that required matching self-raised funds. Both the prize and the first paying customer arrived in October 2021, the point where the company stopped being purely self-funded groundwork.
Seed funding, Azure, and a 146,000% usage spike
2021-10 → 2022-12With a paying customer secured, TelemetryDeck raised a six-figure seed round from 6 private angel investors in October 2022, giving up about 3.6% equity in total [S6]. The founders were explicit that they disclosed this figure to signal they weren't chasing a 'revenue-driven rampage.' The same year they moved part of their infrastructure to Azure under Microsoft for Startups, open-sourced their iOS/macOS SDKs to build developer trust, hired their first non-founder employee for docs and the website, launched their own Mastodon instance as Twitter grew unstable, and introduced a custom query language [S4][S5]. By year-end, usage had exploded: 577 million signals sent to their ingestion servers, a 146,322% increase over 2021, spread across 1,125 organizations and 6.8 million tracked users [S7].
The iOS survey becomes a press magnet while conferences quietly get cut
2022-12 → 2025-10Starting around November 2022, TelemetryDeck began publishing iOS version market-share charts, which grew from blog posts into dedicated landing pages [S11]. By 2025-2026, journalists at 9to5Mac, Forbes, iOS Dev Weekly, and the Wall Street Journal were citing those pages as a go-to source for iOS adoption data. Alongside that, the team tracked its channels directly: social media overall, with Mastodon sending the most website visitors of any social network, was their single most important channel, and Reddit — monitored through a custom crawler called Redreach — stayed a consistent top-10 referrer with no paid promotion [S8]. By October 2025 the company had grown to roughly 4,000 accounts processing over 11 billion data points, adding about 120 new customers a month, while narrowly missing €100,000 in annual sales for 2024. That update also named a channel the founders had tried and dropped: conferences and trade shows, which had 'almost zero impact' on revenue despite eating up large amounts of time [S9].
GEO pivot: AI recommendations overtake social media
2025-10 → 2026-05In September 2025, Lisa began reworking TelemetryDeck's website for Generative Engine Optimization (GEO) — answering the kind of plain questions an AI model might ask on a user's behalf, and adding structured FAQ content [S12]. In July 2025, two years of convertible loans from business angels converted into real voting shares under German law, leaving the founders with a clean cap table and roughly 90% ownership [S9]. By early 2026, new-customer growth had accelerated to 6-13% a month, forcing the team to temporarily cut free-tier data processing down to once a day just to keep up [S11][S12]. The clearest sign the mechanism had shifted: by May 2026 the founders reported that new-customer recommendations no longer came primarily from social media, but from LLMs, with Claude Code and ChatGPT named as the most important AI systems recommending them. The company also finished moving hosting entirely to Hetzner and reported being 'just about profitable' for the first time [S12].
Milestones
- 2020-04Idea stage — Daniel starts building TelemetryDeck during Covid lockdown (date estimated from later '2.5 years' and '5 years' references)
- 2021-02First blog post published; company at the notary for registration; first helper hired for the Viewer app
- 2021-09Rebrand to 'TelemetryDeck'; company officially registering; first paying customers arriving
- 2021-10Won the €36K Bavarian Start?Zuschuss! founders prize (required matching self-raised funds); first paying customer confirmed
- 2022-09Summer Update: partial move to Azure (Microsoft for Startups); open-sourced iOS/macOS apps; hired Marina for website/docs
- 2022-10Seed round: six-figure sum from 6 private angel investors (Encourage Ventures, VDU networks) for ~3.6% equity
- 2022-11Fall Update: state funding received; launched own Mastodon instance; introduced custom Telemetry Query Language
- 2022-12Year in Review: 577,461,576 signals sent in 2022 (+146,322% YoY) across 1,125 organizations
- 2024-128333Narrowly missed €100K in annual sales for 2024 (~€8.3K MRR equivalent, converted from annual figure)
- 2025-07Convertible loans from business angels (raised over the prior two years) converted to voting shares under German law; founders retain ~90% of the company
- 2025-10Fall Update: ~4,000 accounts, 11B+ data points processed, ~120 new customers/month, ~3% average monthly churn
- 2026-02Winter Update: churn ranging 0-6%/month; new hires including first salesperson; iOS survey cited by 9to5Mac, Forbes, iOS Dev Weekly, and the Wall Street Journal
- 2026-05Spring Update: 'just about profitable' and no longer reliant on additional investors; new-customer growth accelerated to 6-13%/month; fully moved to Hetzner; LLM recommendations overtook social media as the top acquisition driver
Whether it fits you
TelemetryDeck's loop depends on structural conditions that don't transfer to every product. Run it if these hold for you; the costs are real and easy to underestimate.
What it needs
You need a data byproduct valuable enough that journalists want to cite it for free
The iOS market-share survey worked because TelemetryDeck's core product already generated this data as a byproduct — publishing it cost little marginal effort but filled a real information gap [S11]. A product that doesn't naturally produce a citable public asset doesn't get this press channel for free.
You can survive 18 months with no paying customers and no outside funding
TelemetryDeck was founded in spring 2020 during Covid lockdown and didn't land its first paying customer or its first outside funding (a Bavarian founders' prize) until October 2021 [S3]. Teams under investor or runway pressure typically can't absorb that long a quiet period before any revenue signal appears.
Founders need to be personally, continuously active in niche developer communities
Mastodon and Reddit only worked as channels because the team was actually present — tracking mentions, replying, engaging — not just posting occasionally [S8]. This only scales if someone on the founding team treats community presence as an ongoing job, not a launch-week task.
What it costs
Maintaining a free public data asset costs real infrastructure with no direct monetization
As new-customer growth accelerated to 6-13% a month in 2026, the company had to temporarily cut free-tier data processing down to once a day just to keep servers running — the free asset that built the reputation also strains the same infrastructure that serves paying customers when it scales [S11].
Spreading across many channels means most of them stay unproven or fraying
TelemetryDeck's own channel breakdown calls Twitter/X a 'necessary evil' with declining engagement, describes LinkedIn lead-gen testing as inconclusive, and calls Bluesky traffic low but 'worth supporting' — running this playbook means carrying several channels that may never clearly pay off, not just the one or two that visibly work [S8].
Cutting a channel that 'should' work in B2B requires trusting your own tracked numbers over intuition
Conferences and trade shows are a conventional B2B move, but TelemetryDeck's founders tracked the actual return and found 'almost zero impact' on revenue while draining time and energy from a small team — acting on that data meant giving up a channel that looks obviously right from the outside.
The numbers we could verify
- churn
- 0-6% per month, averaging ~3% (Oct 2025), with 'too expensive' becoming a more frequent cancellation reason over time (S9, S11)
- new customer growth
- 6-13% monthly growth in new customers from the beginning of 2026 (S12)
- new customers absolute
- ~120 new customers/month as of Oct 2025 (S9)
- revenue
- Narrowly missed €100,000 in 2024; easily exceeded €100,000 in 2025; 'just about profitable' by May 2026 (S9, S12)
- active users pct
- ~20% of accounts log in quarterly (Oct 2025) (S9)
- costs
- Running costs go to 'a rackful of servers and a handful of people'; moved from sponsored Azure hosting to Hetzner in April 2026 for cost optimization (S4, S12)
Channels it actually used
- Social media (Mastodon, Reddit, Twitter/X, LinkedIn, Bluesky)
- GEO / LLM recommendations (Claude Code, ChatGPT)
- iOS market-share survey SEO flywheel
- Word-of-mouth / developer recommendations
- Privacy positioning as differentiation
- Conferences and trade shows (tested, then deprioritized)
Our read
The most visible story here is 'privacy-first analytics company,' but privacy positioning reads more like a moat and a values statement than the actual growth engine — the traceable traffic and press citations came from the free iOS market-share survey and from founders being genuinely present on Reddit and Mastodon, not from the privacy pitch itself.
The 2026 shift from social media to LLM recommendations looks less like TelemetryDeck building something new and more like AI search finding an old trail — the same survey pages and forum mentions that fed journalists for years are now the material Claude Code and ChatGPT read by default. That mirrors a pattern already seen elsewhere in this corpus (e.g. Tally's ChatGPT-as-top-channel shift), suggesting this is a broader 2025-2026 phenomenon, not a TelemetryDeck-specific trick.
'Just about profitable' after six years, a state prize, a seed round, and two years of convertible-loan angel funding is a useful corrective to any 'purely organic bootstrap' reading of this story — TelemetryDeck's patient, community-driven growth mechanism was real, but it ran alongside meaningful outside capital, not instead of it.
Sources
- TelemetryDeck Winter Update2021-02-18
- We are now TelemetryDeck app analytics2021-09-01
- TelemetryDeck Fall Update 20212021-10-28
- 2022 Summer Update2022-09-14
- TelemetryDeck Fall Update 20222022-11-08
- TelemetryDeck Receives Seed Capital From Private Investors2022-10-17
- Our Year in Review2022-12-19
- Social media channel breakdown2025-06-25
- TelemetryDeck Fall Update 20252025-10-02
- 2025 Wrapped2025-12-04
- TelemetryDeck Winter Update 20262026-02-06
- TelemetryDeck Spring 2026 Update2026-05-15
