Learn
What is Entity Resolution?
The process of figuring out when two records in different systems refer to the same real-world thing – the same customer, the same product, the same supplier – even when the data does not match exactly.
What it is
Entity resolution is the detective work of matching records that should be linked but are not obvious matches. "Robert Smith" at "123 Collins St" and "Bob Smith" at "123 Collins Street" are probably the same person – but your CRM does not know that. Entity resolution uses techniques like fuzzy string matching, phonetic algorithms (Soundex, Metaphone), address standardisation, and machine learning to identify these matches with confidence scores. Think of it like a maître d' at a restaurant who recognises a regular customer even when they book under a nickname, different phone number, or a colleague's email.
Why it matters for your business
Every company with more than one system has an entity resolution problem – they just might not know it yet. When you merge customer records from your CRM, website, support platform, and accounting system, you inevitably get duplicates and near-matches. Without resolution, you send the same customer three marketing emails, count them as three separate buyers, and miss that they are actually your fifth-largest account. A mid-market healthcare supplier we worked with resolved 15,000 duplicate supplier records down to 9,200 unique entities. This revealed that they were ordering the same product from three "different" suppliers who were actually the same company with different trading names – consolidating saved them $220,000 per year in negotiated volume pricing.
How we approach it
We build resolution pipelines tailored to each entity type. Customer resolution might weight name similarity, address proximity, and email domain. Product resolution might use manufacturer codes and description matching. We assign confidence scores to every match – high-confidence matches merge automatically, medium-confidence matches go to a human review queue, and low-confidence matches stay separate. This balance of automation and human judgement is what separates reliable resolution from a mess of false positives.
Related Services
Key Takeaways
- •Entity resolution identifies when "Bob Smith" and "Robert Smith" are the same customer across your systems.
- •Every company with more than one data system has duplicate entities – the question is how many and how much they cost you.
- •Good resolution uses confidence scores: auto-merge the obvious matches, flag borderline cases for human review.
- •The ROI is often immediate: consolidated supplier records, accurate customer counts, and better volume pricing.