
How egglife Put Six Retailers' Data in Front of AI Without Exposing a Single Retailer's Numbers
At a glance
- Company: egglife foods
- Industry: Consumer packaged goods, egg-based foods sold through national and regional retailers
- Teams using Nera: Sales and category analytics
- The challenge: Understanding performance across retail partners when the data that explains it is too sensitive to combine or hand to an AI tool
- What changed: Sensitive fields were encrypted before analysis, so retailer data could be combined and analyzed with AI without being exposed
- Headline result: Legal review per new data source cut from about 6 weeks to under 1, and time to cross-retailer insight cut from weeks to days
The problem
egglife is one of the fastest-growing brands in refrigerated foods. Its egg white wraps and other egg-based products sell at retailers including Target and Costco, and the company has been named to the Inc. 5000 list of fastest-growing companies in America.
Growth at that pace raises harder questions. Which retail partners are driving it? Where are pricing and promotions paying off, and where aren't they? How does egglife compare with the rest of the category in each channel?
The answers sit in egglife's retail data: sell-through, unit sales, pricing and promotion results from each retail partner. That data is competitively sensitive, and much of it comes with agreements limiting how it can be used and who can see it. In practice the whole dataset was treated as confidential, not just the fields that mattered. Combining it meant weeks of review before any analysis started, the most sensitive fields were often left out, and a public AI tool was never an option.
egglife had plenty of data and could only see part of what it said.
Protecting the fields, not the whole file
Nera let egglife stop asking "can this data be shared?" and start asking "which parts of it are actually sensitive?"
Sensitive fields such as unit sales, pricing and performance measures were encrypted before they left egglife's environment. Non-sensitive metadata such as product attributes, category labels and store formats stayed readable, so data from different retailers could still be aligned and checked for quality.
The analysis ran directly on the encrypted data. The AI model received the questions, never the underlying numbers. Only the final aggregate results were decrypted, and only for egglife's authorized team.
Fields that had always been held back could now be included, because including them no longer meant exposing them. For the first time, egglife compared performance across 6 retail partners in a single analysis and asked follow-up questions in plain language instead of waiting on a new report for each one.
A repeatable process instead of a negotiation per retailer
Before Nera, every new data source started with the same conversation: who gets access, what can they see, who is liable if something leaks. Each one took about 6 weeks of legal and contract review before analysis could begin.
With raw values never exposed, most of that conversation went away. Adding a new retail partner's data became a standard setup, and review time dropped to under a week. Time from new data to usable insight fell by about 75%, from weeks to days.
More data and more questions don't add exposure, so egglife can expand its analysis without expanding its risk.
Where this goes next
egglife can bring in more retail partners as distribution grows, compare promotion and pricing performance across channels side by side, and enter new markets with the same visibility it has at home.
"Whenever two companies have to combine data they don't fully trust each other with, the review takes longer than the analysis," said Rami Akeela, Ph.D., Founder and CEO of Nera Systems. "egglife's review went from six weeks to under one because the numbers never left their control. They didn't have to choose between seeing the whole picture and protecting the data behind it."