Company Object Accuracy - Interim Guidance

We've identified an accuracy issue affecting the company object attached to delivered signals. Until we ship corrected data, the enriched company attributes we layer on top of that object may be unreliable.

To be precise about where the problem does and doesn't live:

• The signal event is fine. The funding round, filing, job posting, post, etc. is not in question.
• Our resolution to a key is fine. For most signal types we reliably resolve each signal to one of two identifiers, the company domain or the company LinkedIn URL, and you can still predictably match on that identifier.
• The enrichment on top of that key is what may be wrong. The company attributes we derived from that identifier (description, industry, and so on) are the unreliable part.

Where the issue may live

Potentially inaccurate company fields: description, industry, employee count, linkedin_url, and company name. In a smaller number of records the company object may resolve to the wrong company entirely.

Affected signal types (any that carry a company object):

SEC filings (including 20-F and Form D), federal contract awards, reports & fundamentals, GitHub, hiring trends, new business formation, patents, podcasts, product reviews, product launches, SEM / website traffic, Glassdoor, website intelligence, Reddit, Twitter, YouTube, conference speaker, and company LinkedIn posts.

Not affected: web intent signals do not carry a company object. Contact-level fields and job-change signals are covered under separate guidance.

Interim workaround: re-enrich from the trusted key

Each signal already carries a reliable identifier you can use to re-enrich the company yourself (through your own enrichment or our enrichCompany endpoint). Do not rely on the delivered company fields above. Instead, key off the field below for each type:

For the domain- and LinkedIn-keyed types, the identifier itself is trustworthy, so re-enriching from it gives you accurate company attributes today.

Form D and conference speaker are the exception: for these two, the identifier itself was resolved upstream from a company name rather than given, so the identity is the thing that may be off. Confirm the company is correct before enriching from it.

What's next

We're rebuilding our enrichment layer and will re-deliver corrected company data. When it's ready, corrections will be opt-in, will keep each record's original signal ID so they update in place rather than duplicate, and will not modify or overwrite any file you've already loaded. We'll share before/after accuracy for your review before anything is delivered.

Questions or want to be first in line for corrected data? Reach out and we'll coordinate directly.