The average consumer in the United States has 1.65 phone numbers. Here is how I crunched the numbers:

Landlines for every 100 U.S. people	38
Population of U.S.	319,000,000
My calculated landlines in the U.S.	121,220,000
Mobile subscribers in U.S. by top 7 carriers	406,375,000
Mobile + landlines divided by population	1.65

This 1.65 figure is valuable for forecasting the impact of duplicate data in a customer relationship management system (CRM) like Salesforce.com.

Why did I calculate this?

Today, in a co-working office called GeniusDen, I met a gentleman whose lead generation agency is tasked with generating leads for a healthcare services company.

He asked if I had any advice about forecasting the business impact of migrating CRM systems from Siebel to Salesforce/Pardot while undertaking new lead generation campaigns. I said, “Yes,” and I will publish my thoughts as blog post here soon. [Update on 3/2/17: the follow-up post is published.]

One the tips in the follow-up post is to forecast the impact of duplicates at the beginning of the project. This produces a number which you can use to adjust expected ROI downward.

Adjusting for “unknown” duplicates

Before you can forecast the impact of duplicates, you will need to estimate how many duplicates you have. You can find the number of known duplicates with software from CRM fusion like DemandTools. This software works by comparing keys like First Name + Last Name + Phone. Let’s say you find 50 duplicates this way.

You get rid of 50 duplicates, but there could still be 32.5 unknown duplicates in your CRM. These duplicates go unidentified and create downstream inefficiencies which reduce ROI.

How did I calculate 32.5 unknown duplicates?

The average U.S. person has 1.65 phone numbers. If you found 50 duplicates in your CRM keyed off First Name + Last Name + Phone then multiply 50 duplicates x 1.65 phone numbers and you wind up with 82.5 potential duplicates in your database. You identify and merge 50 of them, so 82.5 – 50 = 32.5 remain unknown.

Say, 500 total records remain after known duplicates are merged. You divide 32.5 unknown duplicates by 500 total records to get a rate of 6.5%. Inefficiencies caused by these duplicates are likely to have a proportional 6.5% impact on your ROI. This is bad news.

Last Friday, I went to an event hosted by PMI Dallas. We learned from Nishikant Shirpurkar, a 13-year project management veteran, that sharing bad news early is essential for stakeholder management. If you procrastinate, stakeholders will eventually discover the bad news. When they do, it may be too late to reset expectations.

Like all forecasts, this method of calculating unknown duplicates makes assumptions that may not be accurate. Here is one way to test it:

1. De-dupe a CRM on phone number and use this forecasting method
2. Append additional phone numbers to records with a service like D&B
3. Perform your de-dupe again and compare with the forecast from #1

My guess is that #1 and #3 would be close – assuming your data appends are comprehensive.

If you want me to help you de-dupe your CRM then feel free to reach out via any of my social media channels. I am a Product Marketing and Project Management professional in Dallas, TX. I have an extensive background in master data management, de-duplication, and CRM migration.