skip to Main Content
Machines Vs. Humans – What’s Your CRM Data Worth?

Machines vs. Humans – What’s your CRM data worth?

Last week a Marketing VP asked me about adding data to Salesforce using machine software apps vs. humans like Salesbroom. For Salesforce and similar CRM’s, data enrichment really comes down to trust and how much bad data costs you and your company.

If the company’s going to launch a million dollar marketing campaign based on CRM inputs, then mistakes can’t be tolerated. If an outside sales rep is 15 minutes late for a meeting because the Contact address is partially wrong, not good but no big deal – at least to corporate.

How do machines stack up for accuracy?

For putting chunks of contact and reference data in, machines can copy/paste to text fields (like phone numbers, email addresses and body text) and do some parsing but don’t score much higher than 70% accuracy in part because the original data is often wrong (typo’s etc.) or can’t be machine read (e.g. graphics get in the way). As Thomas Redman points out in Can Your Data Be Trusted, “Much, perhaps most, data will not meet the gold standard, so adopt a cautious attitude by doing your own assessment of data quality.”

Also, the more important performance information that must be populated like “sales stage”, “follow-up when” and “activity type” or the more complex “if this then that” requires humans to enter values correctly into a specific series of fields, picklists or lookup relationships.

And CRM is a lot more than data entry. For Salesbroom data stewards, CRM data enrichment work includes following IF-Then workflows, pick lists, task creation, activity logging, reminder setting – work that requires rules and human understanding.

How do you go back and find machine errors?

Another big problem with the 70% machine success rate is no one knows where the 30% errors are. How do you find and fix them?

So over time, a company ends up like every company with bad data that can’t be fully trusted. Even little things like inaccurate street addresses (e.g. Blvd vs. Ave) or correct data placed in the wrong field have major negative impacts (at scale) that cost users real time and money.

I don’t think we have a single customer that didn’t come to Salesbroom needing to cleanup existing CRM records (so yep, we do a whole lot of data cleaning).

The high level answer is that automated tools are valuable for speeding up certain data work and that humans still need to drive the car.

Rich Schulte

Rich is Co-founder of Salesbroom. Connect with him on LinkedIn.

Back To Top