USE CASE:
Sorting a contact list when the job titles are a mess
Everything below is about sorting a contact list when the job titles are a mess, and about which title classification method holds up on a list rather than one record. Mamba Labs builds Apify actors, and Contact Classifier is the one here that does title classification.
Where this starts
You have 4,000 contacts and 2,700 distinct title strings. Chief of Staff, Head of Revenue, Growth Lead, VP Bizops. Filtering on the word director gets you a Director of Facilities and misses the person who signs. Every routing rule and every sequence split depends on getting this right, and the raw string will not do it.
Traditional Search Tools
- A spreadsheet of keyword rules. Contains and does not contain, in order. Fragile, free, and completely transparent when it goes wrong, which counts for a lot.
- Your CRM's picklist. Forces structure at intake and only works when the person entering the record picks honestly. In practice they pick the first option.
- An LLM prompt over the column. Handles the odd ones a rule list cannot. Costs a call per row and gives a different answer on a rerun unless you pin the temperature.
- Build a lookup. Export the distinct titles, map the top 300 by hand, treat the tail as unknown. Two hours, and it covers most of the list because titles follow a power law.
Mamba Labs Apify Actors
These are the actors we built for this job. If yours is not covered here, search from the bar at the top of the page or contact us here and tell us what you need.
- Contact Classifier5 preconfigured tasks on this page's job.
Preconfigured tasks
A task is one of these actors with the inputs already filled in for a specific version of this job. Open one, swap in your own list, and run it without setting anything up from scratch.