Data Fetcher: Stuck at $6,000 MRR for Five Months, Then $23,000 — Just by Changing How the Annual Plan Was Presented

Dollar figures in this article are paired with rough yen equivalents at 1 USD = 150 JPY.

MRR didn’t move for five months. New signups and cancellations exactly canceled each other out, month after month. “Data Fetcher,” an Airtable extension run solo by London-based Andy Cloke, fell into that state starting in August 2023. What pulled it out was a change to how the pricing page presented itself, not a new feature or a new channel.

As of the Starter Story interview, MRR stands at $23,000 (about ¥3.45M), roughly 4x the $6,000 (about ¥900,000) reported at the previous interview. It’s still just him full-time, with a freelance video producer and a part-time developer friend based in South America providing outside support.

The numbers, organized

What it sells

Data Fetcher is an Airtable extension. It pulls data no-code from 40+ external services (Facebook Ads, Google Analytics, OpenAI, and more) with technically inclined users able to configure arbitrary API requests. In short, it’s “a box that pulls outside data into Airtable.”

There’s a free plan that lets users try 90% of the functionality. On this design, Cloke says: “Data Fetcher sits in a fairly price-sensitive market, and we’ve done a good job splitting free vs. paid features”, positioning the free plan as still a major growth driver today. Paid tiers run on subscription with usage-based caps.

As a side product, he also ships a second Airtable extension, “Charts & Reports”, though as discussed below, that one hasn’t grown.

The turning point: December 2023, changing the pricing page’s default tab

The moment the trajectory shifted is clearly identifiable.

From August through December 2023, MRR went flat. In his own words: “every month, I was losing about as many subscribers as I was gaining.” With a hole in the bucket, no amount of new acquisition raises the water level.

What he did in December was only two things. Set a steep 50% discount on the annual plan. And switched the pricing page’s default tab from monthly to annual. As a result, monthly churn dropped from 10-11% to 7%, and five months of stalled growth resumed.

Not a new feature. Not a price increase. Two moves, changing a default display and setting a discount rate, with essentially zero implementation cost.

Why a 3-4 point churn improvement matters this much

A churn improvement of 3-4 points looks small at a glance. But steady-state MRR can be approximated as “new MRR ÷ churn rate” (the following is this article’s own estimate, not a figure the source provided).

Without adding a single dollar of new acquisition, the ceiling you can eventually reach rises by roughly 1.5x. A few points of churn matters more than tens of percent of acquisition effort. The real nature of the stagnation was a hole in the bucket, not weak acquisition.

The mechanism behind why the hole closed can also be explained structurally. Monthly billing gives customers a “will I keep paying next month” decision every single month. In a price-sensitive market, for a tool that people use some months and not others, having that decision arrive every month is itself an entry point for churn. Shift to an annual plan, and that decision only comes once a year. The 50% discount is the price paid to encourage that shift, halving total payment in exchange for cutting the opportunity to leave down to one-twelfth. The default-tab change simply aligned with the behavior that most users just go with whatever’s shown by default, but combined with the discount, that’s what made it work.

The foundation behind the growth: the marketplace and YouTube

The stagnation fix is a churn-side story, but there’s structure on the acquisition side too.

First, there’s the distribution channel of Airtable’s extension marketplace. It puts the product right where someone already using Airtable is standing the moment they think “I want to pull in outside data.” On top of that, letting people try 90% of functionality for free thins out the barrier between evaluation and adoption.

The other is YouTube. Cloke hires a freelance video creator to produce every Data Fetcher channel video. The reasoning is straightforward: “for this topic, YouTube is far less competitive than SEO, and video converts well.” How to use a no-code tool is inherently easier to convey through video than text, and he’s layering that onto a channel where competition is thin. SEO has been deprioritized.

On support, despite a growing base of paying customers, he’s actually cut daily support emails from 3 to 2. That’s a result of fixing the product and documentation rather than adding headcount. When a question comes in about how usage-cap reset cycles work, for instance, that explanation gets added to the product itself.

Cloke describes competition this way: “the basic functionality could be copied by someone else in a few months’ work. But the edge cases we’ve squashed, the five-star reviews, and the customer relationships aren’t as easy to copy.” Data Fetcher was named G2’s “New Product of the Year” and landed its first enterprise contract.

What didn’t work

Failures are disclosed just as specifically.

The second product, “Charts & Reports,” has sat flat at $300 MRR for a year, which he describes as an “unwanted distraction.” His overall verdict: launching a second product too early was a mistake.

In early 2023 he hired a customer support hire for 2 hours a day, but it didn’t work out and he let them go early. “Their communication skills and technical proficiency weren’t at the level needed,” and “there were red flags during the interview process that I ignored,” he reflects.

During the stagnation period, he tried and discarded several approaches before landing on the December annual-plan discount: introducing a free trial that coexisted alongside the free plan, substantially loosening usage caps, and even eliminating pricing tiers altogether. None of them worked. The single move that worked looks simple in hindsight, but there were three failed moves before it, worth keeping in mind.

The biggest risk sits with the platform

This business’s structural weak point is total dependence on Airtable. And changes on Airtable’s side have caused real pain: rate limits were introduced on the API it depends on, extensions were removed from the free plan, and a pricing change meant “to keep using some features, customers now had to pay $54 instead of $24.” The platform’s direction of moving upmarket also runs counter to Data Fetcher’s price-sensitive customer base.

What’s interesting is that Cloke’s response to this risk isn’t “diversify the product”. It’s “diversify the profits.” He channels earnings into things like index funds, absorbing the risk through his own personal asset allocation rather than the business. He’s not considering a sale at this stage of ARR, prioritizing profitability over growth. As he puts it: “it’s a genuinely solid lifestyle business now,” and “I’ve built the perfect job for myself. I love working on it, it’s the most I’ve ever earned, and I get to enjoy the lifestyle benefits too.”

How far can this be copied

What transfers is the instinct to suspect churn first when growth stalls, and how cheap the fix can be. If new signups are up but MRR isn’t moving, the cause is retention, not acquisition. Setting up an annual plan, choosing a discount rate, and changing a pricing page’s default tab, all of these take hours to implement. The ratio of what you learn by testing to what it costs to implement is exceptionally good. Opening up 90% of functionality on a free plan is also worth considering in markets that are price-sensitive and easy to comparison-shop.

What can’t be copied is the extension marketplace distribution channel itself. Because it stands on Airtable’s user base, even a solo operator reaches thousands of users. Selling the same product as an independent standalone web service would mean an order-of-magnitude higher acquisition cost. And that same channel is simultaneously the biggest risk, if the landlord changes pricing or limits, your customers feel it directly.

The timeline deserves an honest look too. From founding in August 2020 to the interview, roughly three and a half years passed, including five months of complete stagnation. The 50% discount looks like a clean, decisive move only because we know the outcome, at the time, there was no guarantee it was the right answer.

Related reading

• Plausible reaches $1M ARR — a comparable structure for growing a small-team SaaS without external funding
• Carrd’s $2M ARR — a similar approach to pricing design and operational efficiency for a solo-run SaaS