OfferFit take its ML algorithms to the next level with Validio

Tuesday, Aug 22, 20233 min read

OfferFit, an automated experimentation platform for marketers, adopts Validio as its Deep Data Observability platform to secure data quality.

OfferFit collects first-party marketing data from its customers and returns personalized machine learning based recommendations for marketing offers.

Each OfferFit customer has a number of data sources and provides up to 300M records per day. Naturally, securing high quality data for both incoming data, and outgoing recommendations is an incredibly important business objective for OfferFit. In short, Validio helps OfferFit:

  • Save time on troubleshooting, since they get instant insight into nature and location of bad data.
  • De-risk their data product since they can trust that no inaccurate decisions will be made due to bad data
  • Enable data producer data quality ownership by automatically informing those who can fix potential data problems
  • Victor Kostyuk, Co-Founder and CTO of OfferFit is thrilled about the platform:

    “Validio stood out as the best fit for our requirements versus other data observability solutions. I'm very impressed with the product and the team, and looking forward to continuing working with them.”

    Great Expectations was too manual and fiddly to set up

    Before Validio, OfferFit used the open source framework Great Expectations (GX). However, as their business grew, the team noticed that GX was manual and very fiddly to set up. Initially, they wouldn’t know the exact data parameters, so the pipeline with GX tests would fail, and they would need to adjust the thresholds. This dynamic incentivized minimal validation, which defeated the purpose entirely. Kristina Brantley, Senior AI Implementation Engineer adds “Due to memory limitations, I couldn’t use [GX] on my largest datasets directly in Google BigQuery.” In addition, validation was slow and took more than two hours per dataset, versus less than a minute with Validio.

    Validio takes OfferFit “to the next level”

    OfferFit tested Validio during a proof-of-value period. The team deployed Validio on some of their most important and largest customer datasets, and checked for a comprehensive list of potential bad data, including whether there are outliers, whether data is fresh, complete, and whether the distribution behaves as expected. The aim was to answer key business questions related to scaling data validation, ensuring OfferFit customers’ success, and enabling data quality ownership.

    For a comprehensive list of Validio validators, please refer to the docs.

    As mentioned, Validio ran on hundreds of millions of rows in under a minute—a significant upgrade versus Great Expectations. After the trial, the OfferFit team concluded that

    “Using Validio on ingoing and outgoing data will help take OfferFit to the next level.”

    Integrating Validio into the OfferFit product via APIs

    Validio helps OfferFit save time, reduce risk, and enable data producer data quality ownership.

    Validio helps OfferFit save time, reduce risk, and enable data producer data quality ownership.

    Today, Offerfit uses Validio to validate thousands of metrics on hundreds of millions of rows of incoming customer data, ML predictions and recommendations. In addition, OfferFit exposes insights regarding data quality to their customers. Avery Bub, Software Engineering Manager mentions,

    “Validio is a key enabler in making our platform even smarter for our customers. Exposing Validio insights in our product has been straightforward with the Validio API.”

    Validio insights exposed in OfferFit’s product.

    Validio insights exposed in OfferFit’s product.