Research

The data behind why we're building this.

We didn't invent the problems EngineFresh is trying to solve. Here's some of the outside research that convinced us they're real, and what each finding means for how we're building the product.

$91.3B
projected size of the cloud data warehouse market by 2034, up from $13.35B in 2025

Warehouses are becoming the default place data lives.

Fortune Business Insights puts the cloud data warehouse market at $13.35 billion in 2025, growing to an estimated $91.33 billion by 2034, a 23.82% compound annual growth rate. That's the underlying bet behind a warehouse-native product: BigQuery, Databricks, and Snowflake aren't a niche destination for data, they're rapidly becoming the default one. A tool that reads data where it already lives has a growing surface area to work with, instead of asking every customer to duplicate data into yet another system.

Source: Fortune Business Insights, Cloud Data Warehouse Market Report

2:1
margin by which alert fatigue outranks every other obstacle to fast incident response

More alerts is not the goal. Fewer, better ones is.

Grafana Labs' third annual observability survey found alert fatigue to be, in their words, the No. 1 obstacle to faster incident response, by almost a 2:1 margin over the next closest answer. The same survey found organizations running an average of 8 separate observability tools, and 37% citing high costs as a top concern. We take this as a caution against our own product, not just a market opportunity: an alerting feature that fires too often, on thresholds nobody tuned, just adds another noisy tool to that pile of 8. The bar for a real-time alert should be that someone actually wants to be woken up by it.

Source: Grafana Labs, The State of Observability 2025

83%
of data teams now call trust in their data a top strategic priority, up from 66% a year earlier

Teams trust their warehouse more than a second copy of it.

dbt Labs' 2026 State of Analytics Engineering report found that trust jumped from a strategic priority for 66% of data teams to 83% in a single year, alongside a separate finding that 71% of respondents are concerned about hallucinated or incorrect data reaching stakeholders. The same report found most teams still spend the bulk of their time maintaining and organizing datasets rather than analyzing them. Both findings point the same direction: every additional copy of your data, in every additional tool, is one more place trust can break down, and one more pipeline someone has to maintain. Querying the warehouse directly doesn't fix data quality by itself, but it does mean there's exactly one source of truth to trust, not two.

Source: dbt Labs, 2026 State of Analytics Engineering Report

25%+
of organizations estimate losing more than $5M a year to poor data quality

Every extra copy is another place data can go stale.

Forrester research, cited by IBM, found that more than a quarter of organizations estimate losing upwards of $5 million a year to poor data quality, with 7% putting the figure at $25 million or more. IBM's own 2025 research found 43% of chief operating officers now name data quality as their most significant data priority, ahead of cost or speed. None of that is about analytics tools specifically, it's a broader warning about what happens when data gets copied, transformed, and re-copied across systems. It's also the strongest argument we know of for querying a warehouse directly instead of exporting from it: a copy can drift from its source silently. A live query can't.

Source: IBM, The True Cost of Poor Data Quality (citing Forrester and IBM Institute for Business Value)

10%
of data practitioners feel confident in AI-generated insights from their current BI tools

People want to see the query, not just the answer.

In Observable's late-2025 survey of data practitioners, only 10% said they were somewhat or very confident in the accuracy of AI-generated insights produced by their current BI tools. The other 90% ranged from neutral to very unconfident. We read that as a preference for tools that show their work: a chart backed by a query you can inspect, against data you already trust, beats a black-box answer with no visible reasoning. It's part of why every chart in EngineFresh is a query against your own warehouse rather than a number we computed somewhere you can't see.

Source: Observable, The State of BI and Analytics in 2026

None of the research above is ours, and none of it was commissioned by us. We're linking to it because it's what actually convinced us this was worth building, and we'd rather you check our reasoning against the source than take our word for it.

Early access

See it for yourself.

Request early access, or try it first in a sandbox with no production data involved.

Request early access