Looker launched as a developer-oriented analytics tool with just ten customers after twelve months of pre-launch selling. It found product/market fit when data teams and cheap cloud warehouses arrived. Google bought it for $2.6 billion in February 2020. In 2026 Google sells it as agentic BI built on a semantic layer, with seats plus Conversational Analytics metered at $3 per million input and $20 per million output tokens.
Key takeaways
- Looker had only ten customers after twelve months of pre-launch selling. The market shifted toward it about five months after launch.
- Google completed its $2.6 billion acquisition of Looker on 13 February 2020.
- Its semantic layer, once a developer feature, is now how AI agents get trusted answers.
- Conversational Analytics is metered at $3 per million input tokens and $20 per million output tokens, with monthly allowances by edition.
- Include a generous allowance and charge only for heavy use when you add AI to a plan.
Twelve months of selling.
Ten customers.
That's where Looker stood when it finally went to market. It wasn't trying to give big companies simple analytics. It was a developer-oriented platform that let engineers work on the back end so everyone else got the data they needed on the front end.
That's not an easy sell in a market that values convenience.
So what happened? The market changed around it. And then, years later, it changed again, in exactly the same direction.
What the book said in 2022
In Chapter Ten I used Looker as the example of product/market fit arriving from outside.
About five months after launch, two things shifted:
- A new generation of data professionals started joining companies, and they wanted control over their data. Looker gave them that in spades.
- Amazon Web Services launched Redshift, a cloud data warehouse. For the first time, smaller businesses could afford warehousing without buying the hardware.
Suddenly Looker's "too technical" product fit.
The lesson I drew: a good idea may be ahead of its time. If you know market forces are moving your way, staying the course may be all you need to do.
What's happened since
Google bought it. Google completed its $2.6 billion acquisition of Looker on 13 February 2020. Looker became part of Google Cloud.
The developer core became the AI core. In 2026 Google markets Looker as "the agentic BI platform built on a trusted semantic layer," powered by Gemini. It has conversational analytics (ask your data a question in plain English), dashboard agents, and a Model Context Protocol (MCP) server so outside AI agents can query the data.
That semantic layer is the same thing that made Looker hard to sell in the early days. It's where you define, once, what "revenue" or "active customer" means. Now it's what stops an AI agent from making up its own definitions.
The pricing followed. Seat licenses now sit alongside AI usage, metered in data tokens:
| Looker Conversational Analytics (Oct 2026) | Price or allowance |
|---|---|
| Input data tokens | $3 per million |
| Output data tokens | $20 per million |
| Standard edition monthly allowance | 60 million input, 1.2 million output |
| Non-production and add-on instances | 6 million input, 0.12 million output |
| At launch | Promotionally uncapped |
What heavy use might cost (illustrative)
Say a Standard customer's team uses 100 million input tokens and 3 million output tokens in a month. If usage above the allowance were billed at the list rates, that's 40 million extra input tokens × $3 = $120, plus 1.8 million extra output tokens × $20 = $36. About $156 on top of seats. Light users pay nothing extra. Check Google's pricing page for how overage actually applies to your contract.
What will AI usage cost you?
Estimate your monthly LLM spend across models with the free AI Token Cost Calculator.
Why did it work?
Product/market fit can come from outside. Looker didn't find fit by changing its product. The market moved toward it twice: first with cloud warehouses, then with AI agents that need trusted data.
The hardest part to copy became the moat. In MOATS terms, this is Toughness: a sustainable competitive advantage that competitors can't copy overnight. Anyone can bolt a chatbot onto a dashboard. Not everyone has a mature semantic layer underneath. Our guide to moats in the AI era explains why this kind of foundation matters more as AI spreads.
Data and APIs pricing, updated. Of the five pricing models in the book, data and APIs is the one AI has changed most. Looker now charges per seat for people and per token for machines. That's hybrid pricing: a base fee plus metered usage.
Generous allowance, then meter. Most customers will never pay for tokens. Only heavy users will. That keeps the headline price familiar while protecting margins.
When is this the wrong lesson?
"Stay the course, the market will come" is dangerous advice if you can't name the shift you're waiting for. Looker could point to real changes. If you're waiting on a vague hope, you're not ahead of your time. You're just early, and early often means out of cash.
And token pricing makes sense when your own costs are tokens. If you're adding AI to a SaaS product or a service business, a simpler allowance (questions per month, reports per month) is easier for customers to understand. See our guide to token-based pricing for when it fits.
What you can steal
- Find the part of your business that's hardest to copy. Your methodology, your templates, your data definitions, your checklists. Make that the center of any AI offer.
- Include a generous AI allowance and charge only for heavy use. Most customers stay on the base price. Your margin is protected from the few who use a lot. Model it with the AI gross margin calculator.
- Launch new usage pricing with a "no caps" period. Build the habit and collect real usage data before you set the meter.
- Name the market shift you're betting on. Write it down. Check every quarter whether it's happening.
What's the "too technical" part of your business that your competitors skip? Could it become your moat?
AI Token Cost CalculatorEstimate monthly LLM API spend across models.Open the free tool →Frequently asked questions
What is a semantic layer?
A semantic layer is a shared set of definitions that sits between raw data and the people (or AI agents) asking questions of it. It says, once, what 'revenue' or 'active customer' means. Looker built its product around this layer, which is why its answers are consistent. In 2026 Google pitches it as the trusted foundation AI agents query.
What are data tokens in Looker pricing?
Data tokens are the units Google uses to meter Looker's Conversational Analytics, the feature that lets people ask questions of their data in plain language. As at October 2026 they're priced at $3 per million input tokens and $20 per million output tokens, with a monthly allowance included in each edition, for example 60 million input and 1.2 million output tokens on Standard.
How much did Google pay for Looker?
Google completed its acquisition of Looker for $2.6 billion on 13 February 2020. Looker became part of Google Cloud. The deal is a good reminder that a product with a slow, difficult start can still become a highly valuable business when the market moves in its direction.
Can a product be ahead of its time and still succeed?
Yes, if you can survive until the market arrives. Looker had only ten customers at launch but kept going, and two shifts (a new generation of data professionals and Amazon's Redshift cloud warehouse) made its approach fit. The risk is running out of money first, so keep costs low while you wait and watch for the signals.
Why did Google uncap usage at launch?
According to Google's pricing page, Conversational Analytics usage was promotionally uncapped at launch. Letting customers use a new feature freely builds the habit before the meter starts, and gives the seller real usage data to set fair allowances. It's a sensible way to introduce usage pricing to customers used to paying for seats.
Get both books free
The Subscription Playbook (ebook + audiobook) and Feed a Starving Crowd, instantly. In Australia? Claim the free hardcover.
Robert Coorey, MBA
Robert is the author of The Subscription Playbook and Feed a Starving Crowd and a Director at BPAG Accountants in Sydney. More about Robert.
Sources
- Google Cloud blog, Google completes Looker acquisition (13 Feb 2020)
- VentureBeat, Google completes $2.6 billion Looker acquisition (13 Feb 2020)
- Google Cloud, Looker product page (accessed Oct 2026)
- Google Cloud, Looker pricing (accessed Oct 2026)
Facts and prices checked 27 October 2026. Spotted something out of date? Tell us.