One warehouse for ads, orders and CRM,
and an AI analyst who answers in numbers
Most businesses have the data but no picture. Ads live in one dashboard, orders in another, the CRM in a third, and each tells a different story. We put it all in one database and make the numbers agree. Then you get an analyst, human and AI, who answers with SQL queries instead of opinions.
Where your data comes from
We pull from ad accounts: Meta Ads, Google Ads, TikTok Ads and TikTok Shop. From sales channels: LINE OA, Shopify, WooCommerce, custom stores, Telegram order channels and marketplaces. From your CRM and sheets: amoCRM, KeyCRM, HubSpot, Google Sheets. Plus GA4 and PostHog for the site, and your ERP for the real cost of goods. Every source keeps a raw copy and syncs daily, past data included, so history never gets lost.
What lands on your desk
One order book that agrees with your bank, not with what an ad platform claims. In one project we rebuilt a sales channel’s whole order history from a chat log. All 571 of 571 orders came back and matched the bank. The analytics platform had never shown that picture.
Dashboards in money. Spend, revenue, ROAS, CAC and LTV. Retention and how often people buy again. Economics by product and by bundle. Once, an analysis of what people buy together showed that family-size carts brought a far bigger share of revenue than of orders. That changed the whole bundle strategy.
An AI analyst in Telegram. You type a question like “top three cities by revenue last month”. Or “which campaign brought repeat buyers”. Or “margin by product line with real cost of goods”. It answers with numbers from your tables and shows the SQL.
Monitoring. We watch competitor ads and prices, marketplace listings, reviews and search trends. We flag when someone undercuts your prices. In one deployment it caught 184 of 218 such events with zero false alarms. Collectors take only public, non-personal data and respect each source’s terms. They never collect behind logins and use official APIs where available. Alerts land in Telegram.
Unit economics and research. Models for CAC, LTV, churn and payback, built on your real data. Market sizing and competitor teardowns, run by research agents and checked by a person. We have built these for a subscription app and a supplements brand.
The five steps
- Audit. What exists, what gets tracked, what is wrong. One week.
- Questions first. We list the questions the business must answer. The tables follow from the questions, not from the sources.
- Connectors and sync. Raw copies, daily jobs, past data and checks that the numbers match.
- Dashboards and alerts. Built together with the people who will use them.
- AI analyst and monitoring. Read-only access and safety limits on every query. We test it on real past questions before it goes live.
What it runs on
Data and code: PostgreSQL 16, Python, FastAPI, APScheduler. Sources: platform APIs and dbt-style SQL tables. Screens: Metabase or custom Next.js panels, PostHog and GA4. The analyst: Claude and the Telegram Bot API. Hosting: Docker with daily backups.
Packages and pricing
Find out what your analytics actually sees and what it misses. You get a list of fixes.
- Audit of Pixel, GA4, CAPI (server-side tracking) and UTM tags
- Orders checked against platform numbers
- Fix list for your developer or for us
- Dashboard of the gaps
Launch price until 31 March 2027 · from 1 April from $3,600
+ $1,000 with full control: access and instructions handed to youOne database for ads, orders, CRM and site events. Dashboards that finally agree with each other.
- Meta, Google, TikTok and Shopify connectors plus WooCommerce, your CRM, GA4 and PostHog
- Daily sync that also pulls past data
- Ready tables for orders, spend, cohorts and products
- Dashboards: ROAS, CAC, LTV, retention
- Alerts to Telegram
- Documentation and handover
Launch price until 31 March 2027 · from 1 April from $7,100
+ $1,000 with full control: access and instructions handed to youAsk questions in Telegram and get answers with numbers. Plus monitoring of competitors and prices.
- Everything in Warehouse
- AI analyst that can only read data, with safety limits on every query
- Collectors for competitor ads, prices and reviews
- Alerts on price undercuts and unusual spikes
- Analysis of what sells together and by segment
- Monthly review with a human analyst
Prices are starting points for a typical scope. You get the exact number in a written plan after a 30-minute call. Once agreed, it does not change.
Cheaper together
A site that sells + Ad audit and launch + Tracking audit
Launch price until 31 March 2027 · from 1 April from $4,100
3-4 weeks Discuss this packageStore build + Tracking audit + Ad launch + Creative pack
Launch price until 31 March 2027 · from 1 April from $9,300
8-10 weeks Discuss this packageWarehouse and dashboards + Tracking audit + Ad launch
Launch price until 31 March 2027 · from 1 April from $5,400
4-7 weeks Discuss this packageWe build every system in clear modules, with docs and tests. Add full control and it becomes truly yours. For $1,000 on top of any service, we transfer every account and server to you. We also write down how to run the process without us.
- Flexible. The system is built in modules, so it grows with your business instead of being rebuilt.
- You are in control. You run it yourself, with an AI assistant if you like, without waiting on us.
- Any developer can take over. Another team picks it up from the docs in days, so you are never locked in.
- We stay only if you want. After handover there is no monthly fee: you pay for each change you order, and nothing else.
FAQ
How much does an analytics setup cost?
A tracking audit starts at $800. A full warehouse with dashboards starts at $2,500. Adding an AI analyst and competitor monitoring starts at $5,000. The $2,500 and $5,000 tiers are launch prices, valid until 31 March 2027.
How long until we have working dashboards?
Three to six weeks for a warehouse and dashboards, start to finish. The audit alone, if that is all you need, takes about a week.
Can the AI analyst make up numbers?
It can only read from a prepared set of tables, one query at a time, with hard limits on size and run time. It also shows you the SQL it ran. The answer is only as good as the tables underneath, which is why we build those first.
Who owns the warehouse and the data in it?
You do. It runs on your own database, under your account, with your own backups. If you stop working with us, the warehouse and every table in it stay exactly where they are.
We are a small business - is this overkill?
Start with the tracking audit alone. A shop with a few hundred orders a month usually earns that fee back from the audit itself. It finds the orders your current tools do not see.