Skip to content
MakerGain
← Back to project

Formula Bot Becomes Better Analyst: What It Wants to Take Over Is the Part of the Weekly Report Nobody Wants to Do

Edited by MakerGain

Formula Bot has been renamed Better Analyst. It takes over the part of a channel weekly report that refreshes connected data and resends dashboards and reports on a schedule; deciding which numbers to look at and explaining why they changed still falls to a person. Founder David Bressler started it during paternity leave for his second child and kept it as a side project until revenue exceeded his full-time salary, then went full time.

For example, a person has to submit a channel weekly report every Monday morning. They open Google Analytics, Search Console, and Google Ads, pull out last week's numbers separately, put them into one table, make two charts, write three lines of conclusions, and send it to their boss. These steps look simple, but each one requires a person to watch over it, and next Monday they have to do it all over again.

Formula Bot is now called Better Analyst. The official renaming announcement says it is the same account, the same work, and no migration is required; the new name brings connecting, preparing and enriching data, Q&A, dashboards, reports, and automation of repetitive data work into one analysis flow (official renaming announcement[1], read on 2026-11-14). What it wants to take over is exactly the part of this weekly report that involves moving and redrawing things.

In the announcement, the part closest to this weekly report is “scheduled”: Scheduled dashboards & reports can save repeated analyses, refresh connected data, run daily, weekly, or monthly, and analyze new data in connectors on each run; the announcement says updated reports can be sent to an inbox. In other words, refreshing connected data, rerunning saved analyses, and sending out updated dashboards or reports can be handed to a schedule; the remaining “write three lines of conclusions” depends on whether the insights it provides are good enough.

Which Part It Takes Over, and Which Part It Leaves

To get it running, the first step is to connect the sources. This weekly report needs to look at Google Analytics, Search Console, and Google Ads at the same time, and all three sources are on the official connector list; but the plan documentation as of 2026-11-14 says Free gives only 1 data source, while Pro and above are unlimited (Pricing and plans[2]). So what the free plan can verify is “one source, one task”; to connect three sources at once and refresh daily, you have to start from Pro.

The heavier layer of work is called Data Flows, which can turn the moving of data into automated workflows. Pro's fastest cadence is daily, and Team's fastest cadence is hourly; the free plan has no active workflows that can be saved, only 10 successful one-off Sheets tests within a rolling 30 days. The trade-off is put quite plainly: you have to pay first before daily or hourly runs are possible.

The part that still belongs to people after the moving is also clear: who sets the definitions, who renews a data source authorization when it expires, and who explains the numbers when a report is sent out and they are questioned. What the product takes over is moving and redrawing; “whether this number is right” is given back to people.

Starting as a Side Project During Paternity Leave

David Bressler has worked in data analytics for 15 years. He started Formula Bot during paternity leave for his second child, while still working full time as Senior Director of Analytics at a media analytics firm, and ran the project as a side project until revenue exceeded his full-time salary, when he went full time (Indie Hackers interview[3]).

Image: Source Indie Hackers · original interview photo of the founder[4], obtained on 2026-11-08, original image retained.

In the interview, he describes the product as a tool for people to process data faster whether or not they have a technical or analytical background. It supports uploading Excel or PDF, or connecting third-party sources such as Google Sheets and Google Analytics, and then chatting with the data to produce charts, tables, insights, and models, and it also does comment sentiment analysis, classification, and cleaning dirty data. Formulas have not disappeared: in the official credits documentation, chat, formulas, charts, spreadsheets, and dashboards are listed side by side, and none of them consume credits (credits explanation[5]). Formulas are still there, just placed on a longer chain.

The plan tiers also map onto the scenario of one person doing a weekly report. Pro's positioning copy says “for one person doing real work,” and the plans page suggests this tier for people who “use it every week to do their own work”; Team, meanwhile, lets multiple people share the same set of data (Pricing and plans[2]). This is a tiering of plan capabilities, not about which tier is sold first.

The title and opening of that interview say he organically grew Formula Bot to $500k ARR; the title and opening give no cutoff date or currency code. This figure is not profit, nor is it cash received for the full year—officially, model invocation, hosting, payment, and support costs remain undisclosed, and the profit margin remains unclear.

For One Person Doing a Weekly Report, Which Tier Do You Buy?

Back to the weekly report from the beginning. Free is $0/month and gives 1 data source, 1 dashboard (manual refresh up to 5 times a day), and 3 saved tasks that can only be run manually—enough to verify “whether it can understand this data of mine,” but not enough to save Monday morning. To connect all three sources—Google Analytics, Search Console, and Google Ads—and refresh daily, Pro is $30/month ($240 charged once for annual payment, equivalent to $20/month on the page): unlimited data sources, daily scheduled dashboard refresh, and scheduled email reports and daily Data Flows are also unlocked.

Team is $149/month for 5 people, and what it adds is the ability for multiple people to share the same set of data: shared workspaces, Slack delivery, row-, table-, and column-level data access rules by member, and hourly Data Flows; viewers are free and do not count toward the 5-person limit. What it changes is how the same report is sent out with per-person permissions, not whether the report itself still needs a person to set the definitions.

What really speeds up consumption is batch work. AI credits are used only for batch work such as PDF analysis/conversion, enrichment, text analysis, and deep research; chat, formulas, charts, and dashboards do not consume credits; PDF analysis/conversion costs 10 per page, and table extraction costs 1 per file (credits explanation[5]). Each scheduled enrichment run deducts again based on the number of rows processed—automation has an ongoing cost here. If credits run short, you can top up at $10/1,000; plan allowances reset on the 1st of each month UTC and do not roll over; when they run out, batch work pauses, paid chat continues as usual, and no overage fee is charged.

There are others on the same path. Julius's official article[6] (marked as last updated 2025-09-10) says Scheduled Runs can automatically run analyses on a schedule, pull the latest data from PostgreSQL, BigQuery, and Snowflake, and send results to email or Slack. What the two have in common is the same action: set it once, run it on a schedule, and deliver the results to people. But Julius's pricing page[7] sells a different set of credits and tiers, credits on the two sides cannot be directly converted, and there is no like-for-like comparison of delivery permissions for scheduled tasks, so there is no basis right now for saying one completely replaces the other.

For the person who submits a weekly report every Monday, the name change will not submit the weekly report for them. It has folded the weekly repetitive refreshing and sending into a schedule; what really remains for people to do is decide which numbers this weekly report should look at, and how to explain why the numbers changed.

Continue reading

AI productivity
The day the accountant came asking for the bookkeeping records, he decided to let AI go through his own inbox

In March 2023, software company founder Romeo Bellon was overdue on bookkeeping and his accountant pressed him for records. He developed Receiptor AI, which scans inboxes, extracts receipts and invoices, and pushes them to accounting software; he launched it with Luigi Fernandez Ortega in August. It turns historical email into a retroactive scan priced by email volume, while subscriptions process only new email; its privacy policy says documents are retained indefinitely and automatic deletion is still in development.

AI productivity
He Lost Half His Customers and Was Happier: BoltAI Raised Its Price from \$19 to \$79

In September 2024, BoltAI founder Daniel Nguyen looked back at raising the price from $19 step by step to $79: customers were down to about half, monthly revenue was slightly below the $39 tier, but refunds and support inquiries fell, and he counted founder happiness in the pricing math. BoltAI is a native Mac AI workspace that brings together multiple providers and local models; what makes this trade-off notable is that it exchanges fewer customers for a lighter solo support burden.

AI audioAI productivityAI search
You Ask Claude What the Client Said Last Week, and It Digs Through Your Voice Notes

Speechyou connects voice notes and meeting transcripts into Claude, letting users search, quote, and jump back to the original audio; the connection uses OAuth 2.1 and is read-only. The annual prices on the two pages are $67 and $100, while monthly is $15 on both. The founder self-reports MRR 1800$; the TrustMRR record shows $1,351 and 104 subscriptions, and the one-person team has listed the product for sale at $35,000.

AI developmentAI productivity
AI2SQL: AI Writes the SQL; You Press Run on the Read-Only Connection

AI2SQL turns natural language into SQL based on real table structures and executes it on a read-only connection; founder Mustafa Ergisi chose to bootstrap, grew by contacting free users one by one and through SEO and word of mouth, and admits he learned costly lessons on retention and pricing. The product sells on the read-only boundaries of sqlGuard and MCP, with three tiers priced at $9, $19, and $39, but the TrustMRR page fields and the founder’s message give two sets of MRR.