Back office ops · Production

SpotOn reduces time to actionable insights by 6x with Snowflake, dbt Cloud, and Metaplane

The problem

SpotOn's Postgres database was chronically undersized and slow, causing ETL job failures, delayed BI reporting, and making advanced analytics impossible. Only two engineers could contribute to data modeling, the team could not keep pace with data requests, and the data stack became a bottleneck as the company scaled.

First attempt

SpotOn's custom Python/Airflow pipelines required advanced engineering skills and were too slow and fragile: backfilling historical data in Postgres took days or a week, models delivered with no quality guarantee had to be rebuilt from scratch if requirements changed, and Postgres could not handle large-scale aggregations or geospatial cross-joins.

Workflow diagram · grounded in source
1
Transactional data ingested
integration
“the data team ingests transactional data from payment processor partners”
2
Snowpipe loads external datasets
integration
“the SpotOn data team used Snowpipe to ingest a weather data set from OpenWeather to augment order data”
3
dbt Cloud transforms analytics models
integration
“financial, operation, sales, and product analytic data were all transformed using dbt”
4
Metaplane ML monitors data quality
ai_action
“Metaplane's machine-learning-based testing approach and ability to automatically add hundreds of tests, they saved engineering time and always received context about potential root causes and downstream impact when data incidents arose”
5
Data incident alert issued
output
“proactively catching data incidents like these, Metaplane helped Ben's team get in front of any issues that would impact downstream stakeholders”
6
Team validates fix before resuming
human_review
“Ben and his team could pull back scheduled reports until they verified that the data was fixed after an issue”
Reported outcome

SpotOn achieved a 600% decrease in time to actionable insights, an 8x increase in engineering output, and $110,500 in savings.
Data model contributors grew 750% from two to 17, model creation time dropped from days to hours, and time to identify data quality issues dropped from hours or days to seconds.

Reported metrics
Time to actionable insights600%
Engineering output8x
Cost savings$110,500
Data model contributors growth750%
Show all 10 reported metrics
time to actionable insights600%
engineering output8x
cost savings$110,500
data model contributors growth750%
ETL contributors counttwo to 17
engineering hours saved per weekat least 10 hours every week
data model creation timefrom days to hours
time to identify data quality issuesfrom hours or days to seconds
engineering contribution (dbt section header)8.5x
engineering contribution percent (results section)750%
Reported stack
Snowflakedbt CloudMetaplaneSnowpipeAirflowPostgresHeapMetabaseOpenWeatherFivetranMeltanoTableau
Source
https://www.getdbt.com/case-studies/spoton
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

SpotOn achieved a 600% decrease in time to actionable insights, an 8x increase in engineering output, and $110,500 in savings.

What tools did this team use?

Snowflake, dbt Cloud, Metaplane, Snowpipe, Airflow, Postgres, Heap, Metabase, OpenWeather, Fivetran.

What results were reported?

Time to actionable insights: 600%; Engineering output: 8x; Cost savings: $110,500; Data model contributors growth: 750% (source-reported, not independently verified).

What failed first in this deployment?

SpotOn's custom Python/Airflow pipelines required advanced engineering skills and were too slow and fragile: backfilling historical data in Postgres took days or a week, models delivered with no quality guarantee had…

How is this back office ops AI workflow structured?

Transactional data ingested → Snowpipe loads external datasets → dbt Cloud transforms analytics models → Metaplane ML monitors data quality → Data incident alert issued → Team validates fix before resuming.

Related back office ops cases