Deduplicating your Salesforce database without Normalizing first is like putting underwear over your pants. It’s the wrong order.
Deduplicating your Salesforce database without Normalizing first is like putting underwear over your pants. It’s the wrong order.
I’ve often talked about the importance of having a Data Plan for a healthy CRM. A data plan is simply a set of standards for CRM data. For example, review the following ways to write:
Director of Human Resources
Director, Human Resources
Director of HR
DIRECTOR OF HUMAN RESOURCES
Human Resources Director
Lack of standards is a problem. Why does it happen? These examples are from a live CRM. What were the sources of data?
Each source provides valuable data, but what is the indirect cost? For example, the data tagged “source= D&B” had the title in ALL CAPS. The data from Broadlook was standardized to “HR Director”. Jigsaw and NetProspex had the title in different formats. The worst case scenario: each source provides titles in different format.
This is a CRM nightmare.
The solution is to provide a system that enforces data standards. Some companies get by with an agreed-upon set of standards, but that is not enough. Enforce is the key word. Having an agreed-upon data standard is a great idea, but in reality you:
So, again, the only answer is *enforced* Data standards. How do you do this?
The only way to accomplish true standards is via technology enforcement. One example of this is Broadlook’s CRM Shield, which simply allows you to pick a set of standards. Once a standard is built, then all data passing through the CRM is standardized.
Recently, one of my partners pointed out that his clients “Don’t want to change the titles”. He believed to save the titles in the way that they were provided. A “Vice President of Sales” did not want to be listed as “VP Sales”. Thus the debate began about Normalization/Standardization of titles.
Now if the client is always right, we want to accomplish this, but what is the true, best solution?
First, there are two types of Normalization, destructive and non-destructive. Destructive means that when the data is changed, something is lost and the change cannot be reverted from. An example would be truncating a title from “Director of Systems Automation” to “Director of Systems”. Once the “automation piece is gone, it cannot be recovered. A non-destructive example would be changing “Vice President of Sales” to “VP Sales”. The litmus test for non-destructive normalization is whether the change can be reverted. Starting from the shortest form, “VP Sales”, normalization rules would allow any variation of the title to be modified and re-modified without any loss.
“If someone lists themselves as ‘Chief Executive Officer’ on their business card or email signature, they don’t want to be listed as ‘CEO’.” This is the counter-argument; to maintain the original state of the data. This is a valid point. Yes, we can change ‘CEO’ to ‘Chief Executive Officer’ and back again, but once it is changed, how do we know what the original title was? The answer is you don’t know.
Personally, I don’t believe in the counter-argument, but my job is not to judge what my clients want, but to deliver results that help them run their businesses more efficiently. I don’t believe in it for two reasons. (1) Really? someone is that vain that they get miffed about their title being correct, but not in the exact format? (yes, opinion only). (2) Non-standard data is a culprit in creating Dupes in the CRM. Dupes are ugly. Dupes costs companies millions of dollars each year in lost efficiency, data cleanup costs, skewed KPI’s and failed analytical reports. You can’t report on non-standard data. Just think about it.
Take my advice and normalize your data. However, if you must retain the original source of titles, you are THAT fickle CRM admin, here is what you should do.
Have 2 title fields in your CRM.
Field 1: The normalized title. Use this for search, KPI’s, reporting and deduping. Having a single standard by which to measure. This will keep your CRM tuned and your sales and marketing teams will love you.
Field 2: The Original title. Use this for display, mailing and client/suspect/client outreach. If you mail, use this field.
Even the Original title should be normalized to a certain degree. Example: “Vice — President of Sales.”. This example has extra spaces, dashes and a period at the end. A simple clean up would yield “Vice President of Sales”.
Regarding Normalization. I am right, the detractors are wrong (especially Gregg) and eventually they will come around to my way of thinking. Until then, I’ve done my job and outlined best practices that can make everyone happy.
Chief*** Executive***Officer!! , Data Guru (yes, Normalization can fix this too)
Harry wanted to update 1.2 million company records with fresh contact data. In the process of understanding his sales process, I discovered that he would be working with about 100,000 accounts per month.
I told Harry “No, I won’t sell you that data”.
The prospect of losing a Fortune 1000 account set my Director of Sales, who was also on the phone, into some deep breathing exercises.
Harry did the typical “but I’m the client”. He ranted, he raved, he cussed, asked to talk to my Manager. I laughed, told him I was the company founder… and he cussed again. I made a joke about him being a Buffalo Bills fan (the Profiler found it in his bio) and we connected. I am a Bills fan too. He got nice and we talked some football. Harry pleaded. “I heard you have the best data.”
“We don’t have any data…who have you been talking to?”, I pressed. Harry told me the referral source. “Yes, what an excellent client example. They are killing it.”, I teased.
“So, Mike bought data from you, but you don’t have any data?”, Harry asked. The question was thick with sarcasm.
“That’s right. You’ve got it. ” I said.
At this point, I think he said something like “Who’s on f*ing first Donato”, through sardonic laughter.
I explained that Broadlook really doesn’t store any data, that we generate it, on demand, from across the Internet, so the information is fresh.
Now, if I kept teasing him (he deserved it), I really would have lost his business, it was time to get serious.
I told him “Selling you data that you won’t use for a year, is a disservice”.
I explained. “You will love me the first 60 days, then data will start to decay. By month 9, I’ll look like every other data vendor. At the end of the 12 months, when it is time for a contract renewal, you will talk to your sales team and they will tell you the data is crap, outdated, inaccurate. You will blame Broadlook and you will not renew”.
“It’s not”, I elaborated. The day I deliver the data, it will be fresh, but if you let it rot, it’s your fault”.
At this point, Harry realized I was looking out for him. Instead of taking a big dump of data that would sit and age inside his CRM, we worked out a subscription plan. 100,000 accounts updated per month. Fresh data every time.
This is the concept of Just-In-Time data. I’ve had many conversations with companies just like Harry’s.
The lesson: Don’t buy data if you are not going to immediately use it. Buy just what you need, when you need it, and no more. Your sales reps will love this decision.
If you’ve heard any of the above comments or something similar; you have a CRM problem. Are their solutions to these problems? Yes, however; today I will give you a deeper understanding of WHY the problem occurs. Many good vendors exist to solve the problems listed above. I want to arm you with a deeper insight, the WHY.
If you understand the WHY, you will be able to:
So what is the WHY?
Contact data decays
If you have a short attention span, if you are brilliant, or have limited reading time, we are done here. That’s all you need and you know what I am going to say in the long answer. Thanks for reading.
First, let’s establish a baseline from the US Department of Labor.
The national average tenure across all jobs in the US is 54 months. That breaks down to 1.85% per month of job attrition. For high-demand IT workers the tenure is shorter with 3% monthly job attrition rate. The rate for Silicon Valley start-ups is almost ridiculous with the average tenure being just over a year.*
*Not from DOL. garnered from several Venture Capital blogs…take it as an extreme example.
A month at a glance does not show the full picture when factored across an entire year. Look at the picture across a year’s time.
While I did my best to outline all the factors, there are some that I probably missed. What I wanted to show is a picture of the WHY data decays. For the past 10 years, I’ve been immersed in data, the acquisition of data, and the analysis of data. From this experience, I know that data decays somewhere in the vicinity of 2.8% to 5.5% per month. This is based on: (1) Hand verified data-sets that I’ve been building and maintaining for 4 years (2) Poll-based responses from 5 years of live training webinars and (3) General industry sentiment. When adding in all factors, the picture changes significantly.
Don’t make the “additive mistake” 5.5% decay per month does not equal 66% over 12 months (it’s actually 36%). You must add each month and it’s data age and then factor in the average data loss across a year. It’s still not a pretty picture. In 2011-2012, the numbers are sitting at the high end of the continuum, somewhere in the 5 – 5.5% range. This means that you can expect data decay to be sitting somewhere around 36% a year.
So what is the health status of your CRM data? Using the Estimator, find out where you stand. Start by finding your CRM Data Update Cycle. This is how often you go through a CRM full data update. Next, based on what you know about the sector you sell into, pick a Monthly CRM Data Decay Rate. As a guideline, in 2011-2012 the average is 4.5-5.5%.
Based on the 2011-2012 numbers, to stay under the 10% threshold, the CRM should be refreshed every 60-90 days.
“Hold on Donato”, I get this statement often, “I’m updating contact information in my CRM all year long”. My response is a question: What percent of all accounts in your CRM do you connect with each year and update their contact info? Typically the response is somewhere in the 15-20% range. Again, you must apply the data decay principal to data updated over a year time. For example, if you connect with 1200 accounts per year, 100 per month, then at the end of the year, the first month’s data is 11 months old, or somewhere in the neighborhood of 33% outdated. If you are doing a fully updating contact information (name, title, email, phone, company URL company phone, etc) , each time you connect with an account, then you can give yourself a 5% data health increase.
This begs a question: Who should be in charge of updating CRM contact data? If you expect sales to do it, I have bridge to sell you in New York. My answer is Marketing or IT, based on how your organization is structured.
Where does this leave us? At this point you should have a solid understanding of the WHY behind data decay. In addition, you should have a more accurate understanding of your CRM Data Health. Lastly, you should want to take action. In this blog, I want to leave it as pure educational vs recommend vendors. Vendors come and go, the percents of data decay may fluctuate, but the concepts should hold true for some time.
There are many products out there that can solve CRM data issues. Solutions include data companies, deduplication vendors, offline data cleanse services, preventative data-technology and data import technologies. Here are a set of questions to prepare yourself with.
I hope you now have a solid understanding of data decay and it implications. Understanding the issue is the first step towards instituting change. If you get one of the vendors on the horn, be nice, you already have answers to questions that they are not asking. They may also be excited to have a well-informed prospect as a customer.
Donato Diorio is a writer, blogger, speaker and the Founder & CEO of Broadlook Technologies, a company that specializes in CRM data health. He can be reached at ddiorio at broadlook dot com.
SIC codes are dinosaurs, they have not been updated since 1987. NAICS codes were updated in 2012, but they are still outdated.
No surprise when you think that these codes are kept updated by the government. Important new technologies are not represented by the SIC and NAICS codes.
What this means is that we must leave the SIC and NAICS codes behind if we are looking to have a clean, updated and current CRM system. What is the solution?
First, ask yourself, where is the best description of what a company does? Simple, the company’s own self-description is the best qualified. So imagine if you could have, live inside your CRM, all the keywords that accurately describe what each of your clients and prospects actually do today.
This technology already exists and it is called Market Mapper from Broadlook.
Market Mapper builds lists of companies, based on how a company self-describes itself. In addition, Market Mapper can be used to segment an existing list of companies.
Here is a live IceBreaker Video that talks about Market Mapper
There is the common, but false belief that a successful CRM implementation is complete once users are up and running and the technology replaces the manual tasks of tracking prospect/customer information. That is certainly a milestone, but not the end of the CRM implementation journey. Invariably, real-world usage leads to requirements not foreseen in initial CRM planning. Salesforce.com pioneered the AppExchange™ and solved the challenge of the second implementation. Thousands of apps, ready with tested business logic, encompassing millions of lines of code can seamlessly plug into Salesforce to complete the second implementation. Before the AppExchange, companies would find themselves with a total re-engineering effort or CRM failure. Well done Salesforce!
Today, the focus now turns to the third implementation; and it is all about data. Unless specifically addressed, every CRM has a problem with dirty data. Regular de-duplication is NOT the solution. De-duplication treats a symptom of the dirty data disease; it does not cure the disease itself. To cure the disease, you must stop dirty data from ever getting into the CRM. Dirty data can enter a CRM via a user, application or the system. Each source of data needs to be addressed.
A CRM Data-Plan is a single-point-of-truth for treating data across an enterprise. It consists of a robust set of rules that details standards for treating data. Amazingly, less than 3% of CRM administrators have a CRM Data-Plan.
The challenge for CRM providers is that they cannot solve the issue with a one-size-fits-all solution. The business requirements of some companies call for verbose data, while others may prefer a high usage of abbreviations.
A good CRM Data-Plan ensures that company names, titles and all forms of CRM data are normalized to a single standard. With a systematically applied Data-Plan, gone are the days of searching for a company in your CRM five different ways. Multiple sales reps no longer work on the same account listed under several different company name formats. Marketing has the ability to segment markets and get a stronger perspective on customers and prospects. CRM data augmentation is no longer a nightmare. With a strong CRM Data-Plan, you no longer have to fear importing new data in your CRM and hearing the word “foobar” from your CRM administrator. The efficiencies gained are tremendous.
Without a method, an unbreakable method to enforce your CRM Data-Plan, the plan is a paper tiger. Broadlook’s CRMShield™ empowers CRM administrators to seamlessly create a CRM Data-Plan and enforce it for users, vendors and administrators.
Using a single global standard crafted by your CRM administrator, CRMShield™ protects and cleans your CRM in real time. Say goodbye to duplicate and dirty data. The CRM Data-Plan is saved in the cloud, at a single location, therefore the entire enterprise, small or large is protected. Changes made to the CRM Data-Plan propagate to the enterprise so all data, even from different silos, conforms to a single standard.
The Dirty Data quiz serves as a report card. How does your company rate? Do you have a CRM Data-Plan? Is your CRM administrator helping or hurting? What are the steps you need to take to keep your CRM clean? Visit: www.broadlook.com/dirtydataquiz
Every company should have a CRM Data-Plan. With this free website, you can get started on the road to a clean CRM system. Visit: crmshield.broadlook.com