Extraction cases, by source name.
One case file per source — the largest system we run against each of them. Each card opens the full file: what the client needed, what the source put in the way, how the pipeline is built and what the numbers are today.
Tesco
2 800 stores, loyalty and shelf priced separately
UK grocery has two prices for a great many products: the shelf price and the loyalty price. Collapsing them into one number, as most feeds do, produces a price index that describes a shopper who does not exist.
- SKU-store rows to date
- 48B
- measured coverage
- 99.0%
Uber Eats
580k merchants, the fee stack per address
Two customers ordering the same dish from the same restaurant pay different totals depending on where they stand.
- menu-item records since 2024
- 23B
- measured coverage
- 98.0%
Turo
420k vehicles, availability by the hour
A car is either available for the window you want or it is not, and its price depends on how long you want it, when you asked, and who is asking.
- availability-days since 2024
- 20B
- measured coverage
- 96.8%
Clients stay anonymous — every engagement is under NDA, and what is published here is what each of them agreed to publish. Under NDA we walk through the architecture, the failure modes and the cost model of any of these against the source you actually care about. Case files for the other three disciplines are in preparation.
Start here
Have a source of your own? Two lines are enough.
Send the source, the fields you need and roughly how often. You get a straight answer within 24 hours: whether it can be done, what makes it hard, what coverage is achievable and roughly what it costs to build and to run. Whatever the size.
Direct
NDA before technical detail, as always. If it isn't our kind of work, we say so in the first reply.