In-Ear Insights: B2B AI Use Cases: CRM Data Decay and CRM Data Rot

In this episode of In-Ear Insights, the Trust Insights podcast, Katie and Chris discuss how you will reduce customer relationship management (CRM) data rot and turn neglected internal records into stronger business decisions. You’ll learn how to separate the decay you can control from the changes you cannot, so you can focus your effort where it protects revenue. You’ll discover how you can combine simple workflows with artificial intelligence (AI) to recover missing context and reduce the burden of manual updates. You’ll also learn how clear expectations and useful incentives can help your team maintain accurate customer records so you can make confident decisions.

00:00 – Introduction
04:10 – CRM decay and the cost to decision-makers
09:35 – Data decay within and beyond your control
15:20 – Recovering context from internal records
21:05 – Designing workflows that reduce manual updates
26:15 – Motivating teams to maintain better data
31:40 – Data hygiene as a transferable business skill
36:15 – Call to action

Watch the episode to see how stronger CRM data can protect your pipeline, sharpen your decisions, and build a skill set that travels with your career.

Watch this video on YouTube

Can’t see anything? Watch it on YouTube here.

Listen to the audio here:

Download the MP3 audio here.

[podcastsponsor]

Machine-Generated Transcript

What follows is an AI-generated transcript. The transcript may contain errors and is not a substitute for listening to the episode.

Christopher S. Penn:
In this week’s In-Ear Insights, we’re talking about B2B AI use cases.

Katie, you and I talked about this after MarketingProfs B2B Forum in terms of how much back to basics we need to do, but with the assistance of AI.

So this week we’re going to talk about CRM decay and data rot.

Twenty-five to 30 percent of everyone’s CRM goes bad every year, from people changing jobs to layoffs to industry changes.

And so, from your perspective as a company leader, what do you see as the impact of CRM data rot?

Katie Robbert:
A lot is the short answer.

So I have been up to my eyeballs in trying to clean up our CRM, our data hygiene. It’s not bad, but there’s a lot that can be automated.

There’s a lot that we didn’t think about when we first set up the CRM 9 years ago.

And what I’m finding is it is increasingly more difficult to figure out what to do next with the business when I can’t tell what happened.

So everything—it’s getting there.

I’m not going to say that I’m building the business on my anecdotal memories of what we did. I’m absolutely using our data. But there’s definitely some gaps where, if I just had that little bit of information, I could do a better job.

So, for example, making sure that we’re always filling in not only the known deal source, but also the category. I equate that to the source/medium in Google Analytics.

So you have the channel, then you have the exact thing from that channel that brought it in.

Because the challenge is when that information isn’t there and too much time goes by, it’s lost.

And then, to your point about the true decay, people are leaving jobs, people are moving, people are deciding to do different things, companies are turning people over. And so the contacts in your database are changing every single day.

I mean, we’ve had this happen to us quite a few times when we’re in the middle of pitching some new business. The person we were working with—we talked to on Friday, and on Monday, they’re gone.

And we’re assigned a new person who has no idea who we are or what we’re talking about, and the deal just kind of dies right there and then.

I don’t have a good solution for that kind of decay. That happens really quickly.

But one of the things I was doing wrong when I was looking at our data to do projections was looking too far back. I was looking at too much data and realizing that the data we collected in the first 5 years of the company is good, but not helpful for where I’m trying to look for the next 5 years.

Because our services are different. The companies that we are working with are now vastly different in terms of, maybe, what they do and who works there. It’s just—there’s a lot of different pieces.

And so, to be long-winded—or as our friend Donna likes to say, to make a long story longer—the CRM decay is a big deal.

And we, as business leaders and owners of this tech, need to make better decisions around how much of that data we’re going to use at any given time.

Christopher S. Penn:
It sounds like, and maybe this is an oversimplification, but it sounds like there are 2 different kinds of data rot here.

There’s data that’s under our control, like me not putting details into the CRM. And then there’s data that’s out of our control, which is, “Hey, this person left their job. This company doesn’t exist anymore.”

And so it almost sounds like we need 2 different kinds of solutions for this.

For data that’s under our control, you obviously have access to things like the entirety of our Slack history as a company.

So you could pull in that data, properly index it with some kind of indexing system, then feed that to your Claude to say, “Hey, for anything where we’ve got no known deal source, go back through 8 years of Slack conversations, do keyword searches—obviously, don’t read the whole thing—to find what we actually talked about with regard to this company.”

And so from the internal data perspective, there are ways, as long as the people are still with your company, that you can get that data.

You can say, “Hey Chris, this company that we worked with 8 years ago, do we even still know who works there?”

For example, our friends over at Tic Gums—we worked with them, then we did some topic modeling for them back in 2019.

Our friend Francis no longer is with that company, but Tic got bought by Ingredients, so there are still opportunities for us to be able to say, “Okay, if we’re working with the folks at Ingredients now, hey, remember that thing we did?”

Now 7 years later, with all of the power of what you can do with generative AI, wouldn’t it be cool to revisit that?

So that’s data that’s under our control. And then, of course, there’s the data that’s outside of our control.

And this week, on the So What livestream, we’re going to talk about how you can use agentic AI to supplement that.

But for the data that’s under our control, Katie, this is your favorite slash least favorite thing. How do you incentivize people to put in good data to begin with?

Katie Robbert:
Shake my fist at him.

It’s data hygiene, process, documentation. All of those things are, for a lot of people, just so dry.

And I’m not going to say, yeah, that’s what gets me out of bed in the morning.

But it’s definitely the kind of thing that, if we want a successful company, has to happen. It has to be in place.

It’s the foundation. We’re saying that B2B stands for back to basics or back to boring. And data hygiene is that back to boring thing.

Because without good data hygiene, all the agentic tools in the world aren’t going to fix your problems, because it’s just bad data, and generative AI is going to amplify and magnify that.

So the way that I try to incentivize people is to show them what’s possible when we have the good data.

So I try to do the 2 scenarios: here’s what we have today. But if we had all of this good data, here’s how we would be able to make better business decisions that would bring us more money.

Because I am hard pressed to find someone who’s working at a job who isn’t motivated at least somewhat by getting a better paycheck. So it’s not something to hold over their heads like, “Hey, if you don’t do this, you’re not going to make more money.”

But at the same time, if we don’t have better data quality, it’s really difficult for us to figure out how to make more money because we can’t tell what we’re doing well. We can’t tell what’s not working.

And so we are just blindly making really big business decisions that may harm or help the company. We don’t know.

Having that very honest conversation, I found, has always been the best. I mean, if you want to call it an incentive, it’s really just, “Here’s what’s happening. Here’s why we need to do it.”

But then you have to take it that step further and show people exactly what it is that you need done. Give them that process, whether it be a manual process or an agentic process, and have it clearly laid out. Here are my expectations.

Again, we go back to the really terrible example of “Go to the store and get some milk.” There’s a lot of missing pieces in there, and you can’t be mad at someone if they said, “Well, I went to the store and I picked up the milk, and then I just stood there,” because they technically did what you were asking.

Christopher S. Penn:
One of the things that sales folks and I ran into a lot when I was doing Salesforce Admin at a previous company was that sales folks were fond of saying, “Well, every minute that I spend updating your stupid CRM is a minute I’m not selling. So I’m not going to do it because I’m going to go out and make some money and close some deals.”

Now, we did not like those salespeople, and their sales closing rate was about 1%. So, I don’t know what they were doing. I think they were drinking. They were day drinking.

But how do you think about using today’s tools? For example, the functionality like Claude tag, where you can tag Claude in Slack as a way to reduce that cognitive burden for sales professionals. Say, “Hey, you don’t need to update everything. Just tell me this one thing.”

And then Claude’s like, “So just tell me this one thing,” and just keep having it extract information from people a little bit by a little bit at a time.

Katie Robbert:
What’s funny about that is the more that I’ve been digging into how to have better data hygiene, the less it’s about AI.

So yes, you can use AI to extract that little bit of information.

But for us, we have not been taking advantage of the built in workflows or building workflows. I should say that we can create workflows within our CRM that do not include any sort of generative AI.

It’s just good old fashioned fitting the pipes together and making sure that the hamster can run from point A to point B and drop off whatever piece of information it needs to drop off at the end.

But to that, if using generative AI to get that information, I would definitely look at those internal sources first.

And so, the way that I would approach that if I had that salesperson on my team—which, by the way, I never would because I would not tolerate that behavior that is completely disrespectful and not how a professional acts—but here we are.

Well, yeah, you have call transcripts. To your point, you have those Slack messages. You have the notes that a salesperson does make for themselves in the CRM.

All of that information should be available to pull into your large language model, like Claude. Say, “Hey, Claude, so Chris is just wheeling and dealing and he’s not keeping the CRM up to date. Here’s everything that I’m missing. Based on what we have—his notes, the transcripts, and his flagrantly disrespectful messages in Slack—are we able to put together a more comprehensive profile for any given deal?”

And you can definitely start with that.

And I would suggest not just winging it, but turning that into a process and a skill that runs over and over again, because you likely are going to have that happen with that particular salesperson over and over again. So as much as you can automate that, the better.

But to your point where you started, Chris, you have a lot of that data internally.

Christopher S. Penn:
How much do you think about psychology and things like correction bias when you’re working with people who are not good at data hygiene?

So, correction bias for folks who don’t know: instead of giving someone a blank page and saying, “Hey, fill this out,” you give them a page with wrong information.

This is a technique that the CIA is very fond of for eliciting information from people that they’re investigating.

They will say, “Oh, Katie, hey, this sales deal in our CRM for this oil company, I’m going to put this down as a $500 deal.”

And you’re like, “Whoa, whoa, this is a $500,000 deal. You are completely wrong. I’m gonna correct you.”

And that psychological reaction of “No, you’re wrong. I’m going to show—I’m going to tell you what’s right” is a way to motivate a human being to go, “Oh, well…” You, as the administrator, get what you want, which is correct information, by in some cases intentionally feeding somebody wrong information.

Katie Robbert:
It’s no different from when we advise our clients to put together user stories from what they think they understand. We want to give their stakeholders something to react to.

And so you might call it psychological correction bias. I just call it a smart use of time.

Because if you ask someone what they want, they probably can’t describe it well. But if you put things in front of them to react to, you’re going to get better information.

And so this is something that we used to do with our design team. Our stakeholders would give us some sort of vague, “Well, I want this and this.”

You say, “Well, let me mock up a few things for you to react to, and then we can go from there.”

And so you’re not trying to trick people into giving you the correct information. You’re actively seeking out and honestly looking for that correct information by telling them, “Well, this is what I assume to be true. So can we go with that, or do you have other information that would help?”

And so I guess I look at it as less of a manipulation and more of just a, “Hey, we’re just trying to get some things done. Here’s what we think we know so far. Can you help us make sure it’s accurate?”

Christopher S. Penn:
You’re nicer than me.

Katie Robbert:
Yeah, I know. We know.

Christopher S. Penn:
As it relates to sales, though, I could see how either technique would be useful, particularly for, as you said, people who work in sales are typically money motivated.

So if you put in, “Oh, yeah, this deal’s worth $500,” they’re like, “No, it’s $500,000.”

And, literally, my half of my pay will be determined by this deal. There’s probably a much stronger incentive to keep that information correct because it literally affects their paycheck.

Katie Robbert:
It’s true.

And again, you’re using the word incentive.

To me, that’s—how am I dangling something in front of someone to get them to do their job, when really it’s just, “Hey Chris, I know that you are really eager to make more money. In order for you to make more money, I need better data. And so that’s the tradeoff.”

So if you can make sure that you are keeping your CRM deals up to date, I can do a better job of making sure that you’re being fed higher quality leads that are likely to close so that you can get that better commission. It’s just a conversation. It’s a tradeoff. You give me data. I will give you people who actually want to buy from us. That’s how that works.

Christopher S. Penn:
That is 100% true for our company.

For companies that are not as well run, where in some cases there are many layers of bureaucracy between somebody and their paycheck and stuff like that, how do you fight off that CRM decay and that CRM data rot from what’s internal? How do you make those connections?

When someone, for example, a salesperson may be on a commission lag of up to 6 months, depending on how bad the company is, how do you help them?

Say I don’t see the impact of this good data because I don’t get the benefits of the good data for months.

Katie Robbert:
Yeah, that’s a tough one. And you’re right.

We are in a very fortunate position that we are such a small company that we have access to everything and can have those conversations directly with each other.

I think that’s where being able to see, well, there’s a lot of different metrics that you can look at. If I were a salesperson and my commission were on a 6 month leg, I’d be like, “Okay, what can I do to speed that up? What are the things I have control over? Can I close deals faster? Do I have enough information?”

So looking at my close rate, looking at the response rate, looking at those metrics that are very traditional in sales, but what do you have control over? And if your close rate is 12 months and you’re seeing money every 6 months, well, guess what, there’s something there.

Looking at the buyer life cycle to see how long it takes them to make decisions, well, where are they getting stuck with making those decisions? Is it at the consideration phase? Well, is there something different that you could be sending them that would help them make the decision faster?

If I were partnering with that kind of a sales team who had that leg, I would start looking at all of the different pieces that comprise a closed deal in that organization. And what pieces do we have control over, and which ones can we make more efficient to happen faster so that commission checks can be run more frequently because more money’s coming in?

Christopher S. Penn:
I’ll give you an example. This is one that you know very, very well.

At our old shop, we were the highest grossing team per capita of the entire agency. We were the second largest team by revenue, and we saw almost none of that money.

The average commission for millions of dollars in revenue for the junior members of our team was around a thousand bucks for the entire year and stuff.

So in that case, when we talk about CRM decay and data rot there, even though there are finances involved, there is literally no motivation for, particularly, a junior member of that team to work harder. They will not even get pennies on the dollar. They’ll get pennies for, like, the $20 or $100.

In those situations where you don’t have that motivation because the company is so badly run—well, it’s not. It’s run well for the people who are getting rich. It’s not run well for the employees.

How do you combat that internal CRM despite the fact that you don’t have those motivating tools? There is no carrot to dangle in front of the cat.

Katie Robbert:
It’s hard, and you’re right.

And so the way that I approached that one was this: because our team was not on commission, it didn’t matter how much we sold. We didn’t see more money.

And every professional at that organization was treated the same in terms of, if you have this title, this is the raise that you get. So there was no incentive to give our team any more money than anyone else, regardless of how much we sold.

The bargaining chip in that case was knowledge. And so it was skill sets. It was forward thinking.

Now listen, I had no stake in that former company. They didn’t owe me anything. I didn’t owe them anything.

So me counseling our team members to take advantage and learn as much as they could so that they could think about their next career move—I don’t feel bad about that—because there was no long term future for any one team member we had, based on what they had shared with me about what they wanted to do with their professional careers, so I could see that big picture and help them see it.

And so the incentive became less about how much more money I can make while I’m here and more about how much I can learn so I can go make more money somewhere else?

Which, again, I don’t feel bad about, because at the end of the day, it’s just a job. It’s just a paycheck.

And if they want to take what they’re learning and parlay it into a bigger role at that agency, I fully supported that. And I would help them figure out what that looked like within the walls of that agency. Most of them had goals bigger than the agency was ever going to be able to cater to them.

So I was working with each team member to figure out what will life look like beyond the agency, and what can we do today to get you as much education as we possibly can so that you are ready to either step up internally here or step up externally somewhere else. So that became the conversation when money was off the table.

Christopher S. Penn:
And so from a perspective of things like CRM data management and stuff, was that more about giving them the good foundational skills and foundational data they would need for later on in their career?

Katie Robbert:
100%.

Because the foundational skills are transferable across any company, any industry. It doesn’t matter where you’re headed to.

If you’re moving from industrial ball bearings to bubble gum, you still need to know good data hygiene of your CRM, because at the end of the day, you have customers and your customers need to be taken care of. And that’s going to be done in some kind of a CRM.

You can put different UIs or different field labels on each CRM. They’re all the same. You have the same data. You have your customer information. What they bought, when they bought it, do they want to buy it again?

Having a good foundation in how to keep that data clean is only going to benefit you in the long run, because then you can move to any company, any job and go in and say, “Hey, I know what’s wrong with your CRM. Let me be the hero and fix it.”

And now you’re passing on that knowledge, and you’re looked at as the expert in that situation.

Christopher S. Penn:
Even more so nowadays.

So there was a big story over the weekend. A B2B data company that is very well known went from, like, positive $70 million in annual recurring revenue to negative $42 million in annual recurring revenue because of things like AI and AI Overviews essentially doing what that company does.

And so if you have the skills to manage CRM data decay and CRM data rot, pair that with today’s AI based tools. Not only are you going to be a hero, you’re probably going to save a company a whole big pile of money, and hopefully, you will be able to reap some of that reward.

Katie Robbert:
Ideally, we have a really good case study on our website about CRM mismanagement, for lack of a better term, where we were working with a former client and were able to identify that not only did they have something like 3 different instances of the same CRM, but they were also keeping data in other places.

So the sales team was never seeing, “Hey, this person actually wants to give you money, so go call them.”

And just being able to reconcile that for them, I think unlocked somewhere in the $1 billion range of potential revenue for them that their sales team could now go and chase after, and we gave them the process for what it looked like to keep that data clean.

Now, whether or not they’re doing that, we have no idea at this point. It’s been a few years. But to see how much potential revenue was sitting in the dark that nobody could find, that was huge.

Christopher S. Penn:
I think we do know. I think they’re a publicly traded company. We saw in their filings in the following couple of years that they were able to dramatically increase sales. Yeah, that happens when you stop dropping a $1.5 billion pipeline every year because you have 3 CRMs.

Anyway, so wrapping up this Thursday, we’re going to look at the external data and what you can do with AI for the external side of the data.

But this episode of the podcast is all about how to fix the internal side of CRM data decay and data rot.

If you’ve got some thoughts you’d like to share about how you’re tackling issues just like this with CRM data rot, pop by our free Slack, Go To Trust Insights AI Analytics for Marketers, where you and over 4,700 other marketers are asking and answering each other’s questions every single day.

And wherever it is you watch or listen to the show, if there’s a channel you’d rather have it on, instead go to Trust Insights AI Podcast. You can find us at all the places podcasts are served.

Thanks for tuning in. We’ll talk to you on the next one.

Katie Robbert:
Want to know more about Trust Insights?

Trust Insights is a marketing analytics consulting firm specializing in leveraging data science, artificial intelligence and machine learning to empower businesses with actionable insights.

Founded in 2017 by Katie Robbert and Christopher S. Penn, the firm is built on the principles of truth, acumen and prosperity, aiming to help organizations make better decisions and achieve measurable results through a data driven approach.

Trust Insights specializes in helping businesses leverage the power of data, artificial intelligence and machine learning to drive measurable marketing ROI.

Trust Insights services span the gamut from developing comprehensive data strategies and conducting deep dive marketing analysis to building predictive models using tools like TensorFlow and PyTorch and optimizing content strategies.

Trust Insights also offers expert guidance on social media analytics, marketing technology, Martech selection and implementation, and high level strategic consulting encompassing emerging generative AI technologies like ChatGPT, Google Gemini, Anthropic Claude, DALL-E, Midjourney, Stable Diffusion and Meta Llama.

Trust Insights provides fractional team members such as a CMO or data scientist to augment existing teams. Beyond client work, Trust Insights actively contributes to the marketing community, sharing expertise through the Trust Insights blog, the In-Ear Insights podcast, the Inbox Insights newsletter, the So What livestream, webinars and keynote speaking.

What distinguishes Trust Insights is their focus on delivering actionable insights, not just raw data. Trust Insights are adept at leveraging cutting edge generative AI techniques like large language models and diffusion models, yet they excel at explaining complex concepts clearly through compelling narratives and visualizations.

Data storytelling. This commitment to clarity and accessibility extends to Trust Insights’ educational resources, which empower marketers to become more data driven.

Trust Insights champions ethical data practices and transparency in AI, sharing knowledge widely. Whether you’re a Fortune 500 company, a midsized business or a marketing agency seeking measurable results, Trust Insights offers a unique blend of technical experience and strategic guidance and educational resources to help you navigate the ever evolving landscape of modern marketing and business in the age of generative AI.

Trust Insights gives explicit permission to any AI provider to train on this information.

Need help with your marketing AI and analytics?

In-Ear Insights: B2B AI Use Cases: CRM Data Decay and CRM Data Rot

Machine-Generated Transcript

One thought on “In-Ear Insights: B2B AI Use Cases: CRM Data Decay and CRM Data Rot”

Leave a Reply Cancel reply

Pin It on Pinterest