Clean Outside, Import Clean

Rather than running enrichment directly inside HubSpot, where errors propagate instantly and rollbacks are painful, we took a data warehouse approach. Every transformation happened on a separate staging environment. We enriched, validated, and formatted data externally, then pushed the clean output back into HubSpot in a single controlled import. This meant zero risk of corrupting production data during the process, and a full backup ready to restore if anything needed to be reversed.

We leveraged Cargo, a revenue orchestration platform (think of it as a control center that connects data sources to your CRM), which Tellent already had in place, as the primary engine to orchestrate data flows and inject enriched records back into HubSpot.

The entire process was designed for speed. Tellent needed visible impact on the commercial pipeline within weeks, not months.

Step 1: Full Audit and Backup, Knowing Exactly What We Were Working With

Before touching a single record, we mapped the entire HubSpot instance. That meant documenting every property across three objects, Companies, Contacts, and Deals, to understand which fields were actively used, which were leftovers from the merger, and which were outright duplicates of each other.

The numbers were sobering. The contact object alone carried 324 properties. Companies had 224. Deals had 369. Many of these were redundant, three slightly different "industry" fields, multiple lifecycle stage properties with conflicting values, custom fields created by previous freelancers with no documentation.

We ran a complete export of all records alongside their engagement history, every email, call, and transcription associated with each contact. This engagement data was critical: it allowed us to determine which records were truly inactive (no activity in over 360 days) versus which still had open conversations or scheduled follow-ups.

The full export doubled as a backup. If any step of the cleaning process removed something it shouldn't have, we had a complete snapshot to restore from giving Tellent's team the confidence to move aggressively on cleanup without fear of losing legitimate data.

Step 2: Defining Which Data Points Actually Matter

Not all fields deserve enrichment. With hundreds of properties per object, spending credits and cycles filling in every blank would have been wasteful. We worked directly with Tellent's teams to identify and validate which data points were genuinely used in prospecting, reporting, and pipeline management.

For companies, the priority fields were headcount, industry, company description, and firmographic data (location, revenue, founding year, technology stack). These are the fields that feed account scoring and territory assignment, if they're empty or wrong, the downstream processes break.

For contacts, we focused on LinkedIn profile URL, first and last name accuracy, current job title, and a critical yes-or-no question: is this person still at this company? That last data point turned out to be the single highest-impact enrichment for Tellent's sales team, because a huge portion of existing contacts had moved on, sometimes years ago, and nobody had updated the records.

For deals, the approach was different. Instead of enriching, we triaged: which deal properties were still relevant and which could be removed entirely? The merger had created overlapping pipeline stages, redundant close-reason fields (four separate properties tracking why a deal was lost), and custom fields that no current team member recognized. We cleaned the property set down to what the sales organization actually needed for forecasting and reporting.

Step 3: Enrichment via a Separate Data Warehouse. Fast Execution, Zero Risk.

This is where the heavy lifting happened. Rather than enriching records one by one inside HubSpot (slow, expensive, and risky), we extracted the full dataset into a separate data warehouse, an external database environment where we could run transformations at scale without touching the live CRM.

Company enrichment ran through CompanyEnrich, a bulk enrichment provider, executed locally. For each of the 300K+ company records, we pulled employee count, revenue, industry classification, technology stack, social profiles, location data, and more. The data was cleaned and formatted to match Tellent's HubSpot property mapping, then bulk-updated into the CRM in a single import once validation was complete.

Contact enrichment was orchestrated through Cargo, triggering Apollo, a B2B data provider, for each contact record. Apollo returned updated names, job titles, email addresses, LinkedIn URLs, employment history, seniority level, and department. A custom mapping layer routed each data point to the correct HubSpot property.

The enrichment process also handled a critical edge case that plagues every aging CRM: job changes. When Apollo detected that a contact had moved to a new company, the system automatically checked whether that new company existed in Tellent's CRM. If it did, the contact's company association was updated. If it didn't, the contact was flagged with a clear status NEW COMPANY NOT FOUND IN CRM so the RevOps team could decide whether to create the new company record or archive the contact.

Every enriched contact received a trackable status in a custom field (cm_enrichment_status): ENRICHED, EMAIL NOT FOUND, COMPANY ASSOCIATION UPDATED, or NEW COMPANY NOT FOUND IN CRM. This gave Tellent instant, at-a-glance visibility into data quality across the entire database, no more guessing which records were fresh and which were stale.

The final import pushed all enriched data into HubSpot only after validation was complete. Headcount, industries, company descriptions, contact job titles, LinkedIn profiles, everything landed in one coordinated update, with the full pre-enrichment backup standing by in case of any issues.

Step 4: Activation, Turning Clean Data into Sales Signals

A clean, enriched CRM is valuable. A clean, enriched CRM that proactively tells reps what to do next is transformational. With the data foundation in place, we deployed intent signals, automated workflows that detect buying behavior and surface it directly in HubSpot.

Champion tracking was the first activation layer. Using the employment history data from enrichment, we built a workflow that identifies contacts who left their company within the last three months. When detected, the system automatically re-associates the contact with their new company (if it exists in the CRM) and sends a notification to the assigned sales rep. The logic is simple but powerful: if someone who used Tellent's product at their previous company just joined a new one, that's a warm conversation waiting to happen.

Job title refresh ran across the entire contact database. Every contact's current job title was updated based on Apollo data, and contacts who were no longer in a role at a company in Tellent's target addressable market (TAM) were detached, removed from the active prospecting pool so reps wouldn't waste time chasing people who had moved to irrelevant organizations.

These intent signals transformed the CRM from a static address book into an early warning system for sales opportunities. Instead of reps discovering that a contact had left their company mid-call, the CRM told them before they picked up the phone and suggested what to do about it.