Learn how your CRM is lying to you and how those lies affect not only your data but also the critical, strategic decisions you're making based on the flawed data in your CRM that impacts every department in your organization.
Prospectify has seen the underbelly of data that powers the world of sales, marketing, and customer success. It’s disturbing.
Our own control group of data shows that once-pristine B2B data (such as data that lives in CRMs, marketing automation platforms and customer success platforms) is on pace for an astounding 54 percent decay rate after one year! That’s data that’s gone untouched the whole time. And what surprised us most is that over 50 percent of this data is attributed to “accept all” or “catch all” email addresses that are not verifiable (but more on that in a moment.)
In one case, we cleaned, deduped, and processed a file of 85 million B2B contact records to our standards, and when we were done, it turned out to be ~16 million records. Yeah, 81 percent of this vendor’s file turned out to be out-of-date, duplicative, or otherwise inaccurate.
We spoke to a Fortune 100 sales director who said, “Once a contact goes in, it never gets touched again.”
He’s not alone. We’ve also spoken to several people working at or around a well-known CRM provider, and they told us that across the board, their customers have issues with decayed data which haven’t been solved.
The first step to solving this problem is understanding how it affects your business. Here are the lies your CRM is telling you.
This data inaccuracy problem gets worse as you scale, as each ill-conceived decision has an exponential impact, not just for your sales team, but across all functional departments of your organization.
Consider all the strategic, tactical, and budgetary decisions that are made based on both realized results and forecasts from the revenue-generating departments in your company, including expansions, new headcount, and product development.
Now, consider, as just one example, the implicit and explicit costs associated with hiring, recruiting, onboarding, training, equipping, etc., each new member of a team if the decision was based on goals determined from insights gleaned from data in your CRM.
Now that you’ve pictured what that looks like, consider that across all CRMs and data sets we’ve assessed, cleaned, and enriched, the average percentage of decayed or otherwise inaccurate data is over 40 percent!
Let that sink in. An average of 40 percent of the data that organizations everywhere are using to generate predictive analytics and insights to make mission-critical decisions … is fundamentally flawed.
As IBM and HBR have illustrated, the so-called “rule of 10” provides a simple means to estimate these costs. It is based on the observation that “it costs 10 times as much to complete a unit of work when the input data are defective as it does when they are perfect.”
Predictive analytics has far-reaching capabilities across industries, with millions of use cases at play. Major corporations in banking, retail, insurance, and manufacturing have been looking at predictive clues in data for decades.
The new buzz around predictive analytics is due to the increase in mainstream applications of machine learning and AI. Many analytics platforms claim to use a machine learning model that makes precise predictions based on incoming data.
There are two types of predictive models: classification and regression.
Classification relates to status. For instance, is a prospect “at risk” (unresponsive, low usage) or "engaged" (responsive, high usage). These statuses are the result of classification models. Regression models, on the other hand, predict a number. There are various forms of regression models (linear, logistic, etc.) that all look at the relationship between variables.
The tech behind these concepts, when applied to a CRM, for example, can do a few things.
As you compile historical data, you can build predictive models around customer actions, and these predictive models can generate recommendations. In sales and marketing, these recommendations are most useful when they work to increase revenue and/or fight churn. This can be done through the data analysis of lost or unsuccessful customers.
In a practical sense, data-powered formulas are created so decisions can be made based on the calculations. Certain factors can be calculated using twice their value (2x) for better accuracy. For instance, if your product prides itself on being easy to use, support ticket count might be doubled in a customer health score model.
For SaaS companies, there are use cases across departments.
Sales reps can identify which prospects are likely to be successful through a mix of firmographic data, technographic checks and traits unique to their product.
Predictive analytics help marketers target communication more precisely by not focusing on those who won’t respond or those who will engage without being contacted first.
Formulas power customer success software platforms like Gainsight and Totango. For instance, “Log-ins + Response Rate” or “Support Ticket # vs. Usage.” Other business intelligence rules can be created, as well. A common example: Once a customer reaches a specific usage milestone, renewal or conversion is more likely.
So, where’s the lying?
Behavioral data collection and analysis relies on data accuracy. However, data decay and inaccuracy is a glaring problem in the B2B world,which predictive analytics tools fail to address. Based on what we've seen at Prospectify, this problem cannot be overstated. It's both implicitly and explicitly costing organizations millions of dollars.
Here are two prominent examples.
Let’s say you create a behavior-triggered campaign that begins an email drip whenever a lead downloads or reads more than two pieces of content. The emails in your drip campaign are sent to prospects over the course of 30 days.
A Prospectify data study showed that 4.5 percent of those emails would bounce. That’s not ideal, but it seems manageable.
But since completing that study, we’ve deployed our new “Deep Verification,” which is part of our Always-On CRM cleanup. Applying our Deep Verification to that sample, we found that another 4-5 percent of those would be sent despite the data being inaccurate — and you likely wouldn’t know it. For example, you wouldn’t get a bounce notification.
This is an expensive problem for teams who make decisions based on data and results. That particular damage is a result of the proliferation of “catch-all” email domains (also commonly referred to as “accept-all” domains).
Catch-all email addresses were created with good intentions. It ensured that no email to the domain would be rejected and lost. Catch-all domains accept all email without rejection. Though useful for those concerned about potentially missing important messages due to typos in the mailbox, spammers soon took advantage of the opportunity this presented.
Today, they’re often implemented as a means to obfuscate the mailboxes that actually receive mail to counteract SPAM by deleting incoming messages when no recipient mailbox is found.
To make matters worse, the decay rate on data compounds, just like a crappy credit card. The aforementioned data study now shows that nearly 18 percent of the randomly sampled contact data is invalid, out of date, and non-confirmable. This occurs after just 90 days.
Now let’s look beyond that one case study. After performing our Deep Verification on Prospectify’s customer data, we found that 44 percent of data in the average CRM is outdated, shows that around. Here’s a sample Summary Data Sheet we offer to customers interested in our Always-On CRM Cleanup. A few vital points:
The adverse effects of this are magnified with each action.
One Prospectify customer studied their sales reps for several weeks and found that, on average, each rep was spending six hours each day researching data, correcting bad data, or other menial tasks that are not high-value sales activities. If that’s even partially true for your sales reps, consider that cost to your company per year.
Here’s a quick visualization of that cost:
Data inaccuracy can also dilute or ruin efforts such as identifying cost-per-action, A/B testing, headcount decisions, and forecasting.
For example, an organization may determine to focus the efforts of their outbound salespeople and their organic and PPC marketing efforts on the company profile that generates the largest contract values.
They analyze their post-sale CRM data and see the profiles include companies in the retail space, built on Magento, that have 200-500 employees. And the main point of contact and decision maker needs to be VP-level or up in the marketing department.
Can you see how it all goes wrong here without a smart, comprehensive data hygiene process?
The regression model, powered by deeply flawed data, and their Total Addressable Market (or, again, what they think it is) may tell them, for example, that they can grow their revenue by 40 percent the following year by executing that plan.
So, let’s examine just some of the costs of executing this plan:
All of this is on top of the already ballooning costs of the existing sales reps not being able to maximize their time spent on high-value sales activities.
We’ve discussed the power and peril of predictive analytics. But how can you determine which is what?
The first step is to be a critical shopper, a skill that is increasing in importance with the growth of SaaS products and tools.
Harry Frankfurt, American philosopher and author of “On Bullshit,” said (paraphrased):
“For the bullshitter, he is neither on the side of the true nor the side of the false. He does not care whether the things he says describe reality correctly. He just picks them out, or makes them up, to suit his purpose.”
Technology vendors and tools may not set out to lie, and may not know the extent of their embellishments. But their purpose is to sell, so the truth might be stretched.
Look for vendors that are offering you value outside of the direct sales process. It’s these organizations that typically are deeply familiar with the challenges you face on a day-to-day basis and aren’t just selling you the all-weather floor mats of your industry.
The New Yorker published, “How to Call B.S. on Big Data: A Practical Guide,” and offered some tips on the matter, a few of which apply to CRM data and analytics:
The second step of the BS test is implementing a a process to identify and clean your data inside your organization. We, of course, recommend our Always-On CRM Cleanup, which includes the aforementioned Deep Verification. Doing this will defend against false positive signals from catch-all domains and provide more precise verification than a simple email verification. This could, in turn, save your organization up to millions of dollars per year while exponentially increasing profitability.
Without a robust data hygiene solution in place, you're not able to maximize your revenue potential. Prospectify's Always-On CRM Cleanup will ensure that your CRM data is accurate and up-to-date.