Episode Transcript
[00:00:00] Speaker A: That contract you meant to cancel, it just renewed. Trailblazer Renewal Tracker reads your PDF and word contracts right onto your Windows PC, pulls out every renewal and deadline date and drops them straight into your calendar. No spreadsheets, no sticky notes, no surprise charges. Track your first contract free forever. Search trailblazer Renewal Tracker in the Microsoft Store and never miss a renewal again.
Hello, and welcome to what Counts. Every organization hides a story in their data. This is the podcast that digs into the governance problems people inherit, ignore, and discover too late. I'm Lee, and as always, this is my co host, Maura Dunn.
So, Maura, last episode ran a little long and we ended on two questions that you actually said out loud as the thing to pick up next. The first one was, how do you determine what the last one was, what the last piece of data was? And the second one was, how do you determine the reliability of your data?
So I figured we'd just walk right into those two today. How do you feel about that?
[00:01:06] Speaker B: So I hate it when you throw my words back up at me.
Those are two important questions, and they followed on from my bold statement at the beginning of our last episode about stop making copies, which I stand by that. I stand by that statement.
Even as I sometimes break the rule, I stand by it.
So what we're trying to get at with all of these questions that we're asking and bold statements is we depend on data, business runs on data, and our technology isn't oriented to make data easier to find, not easier to find, not easier to rely on, not easier to store. It's really good at creating.
Our technology is really good at creating data. It's also very good at sharing data.
[00:02:02] Speaker A: I think the search function works really well. I can't say really well the search function works, but it just brings back everything and you still got to sort through everything, right?
[00:02:14] Speaker B: So now you're, you're talking to the librarian and me about the different types of source of searching you can do. We've all become very accustomed to what a librarian would call universe searching. The keyword searching, where you're searching broad swaths of data, looking for keywords a little bit more sophisticated maybe than the exact matching keywords that it started out, you know, a decade or so ago. But we get the fuzzy logic of, you know, does tree and bush both bring you back a similar result? Because what you really want is plants, maybe.
But all that does is expand your results when you do that kind of universal searching across many, many sources. And our access to sources only grows by the Minute that universal searching is designed to bring you the universe. Librarians are really good at naming things, hence the name.
It brings you everything.
That doesn't help because then you have to sort through it somehow. You have to filter it and find what you're looking for. And I've talked before about how some things are very low stakes. I want to know where the closest gas station is. It tells me there's 10 of them and they're within 10 miles. And the map helps me figure out which one is on the way to where I want to go. It's pretty low stakes because I don't have a strong preference over one gas station for another.
But when you're talking about your business and your critical business information, the stakes are a lot higher. When you're looking at, say, inspection, leak repair, construction, design information for an oil or gas pipeline, or even a water pipeline, which is not volatile.
[00:04:03] Speaker A: How about just your gas health insurance quote that you agreed to?
[00:04:10] Speaker B: Yeah, we'll come back to that one. But the. But in the. Like, the safety aspect of can we do we know what's happening with this pipe?
And you search across everything and you find a lot of stuff about this pipe. But actually knowing what the most recent change was is critically important.
So when it's that high stakes, you don't use a universal search. You go to a database that is built for purpose to hold the data and help you find what happened. Most recently, in many companies that we've seen, that data is spread across multiple databases because you might have a leak database, an inspection database, and a repair database. And you might have the construction and design information in documents somewhere. So, but at least you know, within those four places, you understand how it's organized, you understand how it got there, and, and the work a person has to do to find the right thing is reasonable.
It's not ideal, and I want to talk about that later, but it's reasonable. The problem is that when you're not talking about safety for something like a pipeline, when you're talking about that most recent health insurance bill or health insurance policy or, or the last upgrade you did to your business insurance, your general liability insurance, you don't take the time to build the database that isolates leaks and repairs and inspections and construction because you're like, oh, it's a document. I'll just read it. I don't have that many of them.
[00:05:47] Speaker A: Once a year I'll read it, Right?
[00:05:49] Speaker B: Once a year I'll read it. Except every time that as a consulting firm, every time you Start a new client engagement.
They want to know. And so you have to produce it, and you have to produce it with certain. Within certain parameters. And it's a. And it's a frenzy because you only do it once a year. So you have to go back to the insurance broker and ask them for another certificate. And each company that you're working with wants a different. A slightly different language on the certificate.
And so you have to go searching for the documents.
And if you're lucky, they're in a folder, a folder that you know where it is, maybe doesn't even matter about the dates yet. Just if you're lucky, it's in the folder that says cyber insurance or insurance, with a subfolder that says cyber or even just insurance. And in there, mixed up is your errors and emissions and your cyber and whatever.
If you're unlucky, there's critical things in an email and you didn't take them out because you were in a hurry. And you're like, I'll get it later. If you're really unlucky, your CFO goes on vacation and you have to figure it out by yourself.
[00:07:10] Speaker A: It just had to be in that one precise moment when I was just getting off the plane.
[00:07:16] Speaker B: Exactly. That's the point. The point is everything's easy until it's critical.
Right. If you had an afternoon to go searching for the certificate of insurance or the most recent email exchange with the broker. And sometimes it's not your fault that you've left that email hanging in your email box because the broker never got back to you because they're doing other things and you lose track of it and you don't follow up until it comes up again. And so you always are in a hurry. You're always under the gun when you're trying to find these things. And your process doesn't support that. Your process is where the breakdown is. Because actually holding all your insurance documents in a folder isn't terrible for a small company. You don't have thousands of them that you have to look through. You have five, and you might have a couple of versions of each of the five, and they renew annually. So maybe you have a folder for every year where you do, where you've put, this is what happened this year.
The problem is that your process only works when everything's fine.
Your process doesn't work when you need to produce it in a hurry.
And so when you talk, when you mentioned the two.
The two questions that we kind of left off last year, the first is about version and the second is about reliability, and they're actually the same question because version is part of reliable.
You have to know that this is, this information is accurate, that it accurately reflects the situation today.
You have to know that this information is reliable because it came from, from a trusted source or it came from the business process that produces that data.
There's a system that is monitoring your throughput or your out.
Excuse me, your throughput or your output, and it produces data and that data gets put into a report. Well, how does that data get put into a report? Does. Is it automatic?
Does. Does some process pull the data from the source and land it in.
In another system where you can do analytics? Does it land it in a spreadsheet as protected so it can't be changed? Or does somebody read it and type it in a spreadsheet where it doesn't have any protection in the spreadsheet? And also it's subject to typos.
And even if they do move it correctly, there's no traceability back to the source.
So somebody goes and looks at that report in eight months and then they look at the source and they're like, well, it's different. Well, probably it's different because something changed. That doesn't mean it was wrong in the first place.
But again, your process doesn't support that. The process is too dependent on manual moves and the data is very separated.
And so I think we've talked a lot about ways to tackle this.
And the first kind of round of conversation that you and I had about this was search through stuff and find the right thing and what can you use in the properties or how can you do comparisons?
And to me, that feels like we've already lost the battle. If that's the thing we're trying to fix, we shouldn't be trying to fix a better way to find something.
What we should be trying to fix is a better way to create and store it.
Because if you create and store it in a consistent process, then finding it is easy.
Understanding that it's accurate and reliable is easy because of the way that you created it and stored it, knowing then once you've stored it, how does it get used? How is it allowed to be manipulated? How does it go into an analytics program or, you know, some intelligence tool that helps you learn things from your raw data?
There's rules there. There's always. You have to have rules in order to do the analytics piece.
So the rules give you that sense of accuracy and reliability.
But you have to have. But you start with, how did you create this data. Where was it stored, how and where did it go?
[00:11:45] Speaker A: You're good there.
[00:11:46] Speaker B: You look very thoughtful and I'm wondering if I lost you.
[00:11:49] Speaker A: No, I get it. I mean, it's all process, which goes back to. People have to follow the process though, so. And it's the same thing with typing into a spreadsheet. Somebody could make a mistake, so somebody could not follow the process.
And now you have a problem again,
[00:12:05] Speaker B: somebody could not follow a process and people have to follow process.
That's true.
But there are things that you can set up to support that person who's trying to follow the process.
And they can be fully automated and we'll talk about those in a second. And they can also not be fully automated and you can still provide the backup. So let's talk about the automation, because that's what everybody thinks about first.
There's a lot of workflow engines out there.
Most collaboration tools include kind of basic workflow. You can route something for approval.
If you have an integration from a source system into an analytics system or a reporting system, you build those rules, the computer will follow the process.
If you have defined the data from each system and you've said this data field equals this data field, or the data in this field should go into this new field in this analytics system.
And if we're going to manipulate that in some way, these are the appropriate and allowed ways. So you might wait, weight this piece of data higher than this piece of data, or if you've hit a threshold, then it turns the slide, then it turns the square red or something like that, and. Or it kicks off a notification to somebody. That is a tedious process to write down all those things.
But people know those pieces because we're doing them all instinctively, we're doing them all through repetition and working together.
The act of the art of writing them down, breaking down that process and those rules. It takes time, but once you've done it, then you've set up your automation and it runs for you until you change something and then you go in and you do that step again. What piece changed? What do I need to update?
So it's an investment, it's a huge investment up front in time to do the breakdown and also in software to make it run. So you, you make your choice on investment based on how big you are and how big your risk is.
Without that investment, without that automation investment, with just people trying to follow process, you still need to invest in that first step of what are the, what's the process and what are the rules and you need to highlight a couple of things where you can reinforce your process.
So first thing is do you have a checkpoint? There's a. We used to, when we were, when we were doing this a lot for our clients, we called this record process mapping. Where's the point at which the record comes into the system, it's received from somewhere else or where in the process does the, does the information become a record?
And following the definition of a record, this is capturing a transaction or documenting a right or interest or a process or decision or something.
Follow those rules and identify the, identify those points.
Then you look at your process and we've seen a lot of this over the years of redundant process where you send. I love this one. This is my favorite one always because it seems so simple. You send a resume to the six people who are going to interview somebody and every one of those six people decides to save that resume somewhere.
[00:15:26] Speaker A: And, and yet maybe it's CYA tactic.
[00:15:31] Speaker B: It is because people in the absence of good systems, good process and good guidance are afraid that they will be the one who's lost something right or they're afraid that they'll be on the hook. It really is a very personality driven thing because we have certainly seen people who throw everything away way and whether they should or not, they're just like I got no records.
So in that resume situation the six places. The problem with that is that there are laws around how long you can keep resumes for people who aren't hired and there are potential implications around EEOC requirements and privacy and data breaches.
So you think it's harmless? Oh, I'll just keep an extra copy just in case. But it's not in fact harmless to the organization but comes from a good place. I don't want to be the one who lost it. This might be important so that the. The checkpoint is tell people what to do. The first step, you are being provided this resume because you're going to do an interview. At the end of the interview we want you to send your feedback to this person or this email address or put it into this system. And we want you to destroy the resume because you are not responsible for keeping it. And in fact we don't want you to keep it. It causes more trouble.
Send six copies of the resume in email versus I've posted the resume here. Here's your link to it. That gives you another piece of control that doesn't involve that huge investment in automation. It's typical office software can hold, can hold something in One place and share. It takes an extra step to do that. Email, so much easier.
But that's a checkpoint.
The originator, the owner of this process, which is I'm the recruiter and I have to get people to interview somebody.
That's their job and it's part of their responsibility to make sure that happens.
Then there's a quality, a quality assurance step or a back end kind of check which is periodically you have from a central perspective you go look, are people putting things in the right place?
It can do this.
What's in our SharePoint, what's in our email?
Are people putting things on their hard drives when they shouldn't be, when they should be using their One drives or using the SharePoint site or the Google equivalent of that or whatever version, the Mac equivalent of it, whatever, whatever thing. You've got these layers usually of a hard drive that's physical personalized cloud drive and then a shared cloud drive.
So you can put those checkpoints into even a manual process. And it doesn't have to be onerous, it doesn't have to be 50 checkpoints. It's just we've set it up, we've done the work, we've said for our recruiting process.
We have the recruiting team and their job is to collect requirements for the job and create job descriptions. So they might ask a lot of people for input to the job description. That input is not a record and it should be purged when the job description is created. They own the job description and they're responsible for updating it over time.
And then they go out and look for resumes, they collect resumes, they sort them based on criteria into interview, don't interview.
They share them either through a link or some other way with the interviewers. They collect the feedback and that recruiting team is responsible all the way through for making sure that the data is managed and, and, and honestly, at the end of the day, yeah, you're going to add a little bit up front to do the analysis. You're going to make sure that, yeah, the recruiter has to use a link and not a, not an email so that the reviewers, the interviewers don't have to separately download and then delete their document, their copies of the resume.
But the payoff is you get hacked later and you had 10,000 resumes that people had stored all over the place and things were breached and you have to figure out whose resumes were they and where do you find those people and they shouldn't have been around anymore because your, the EEOC rules and your retention policy both say that you destroy those within one to three years. There's a. There's some changes going on around how long that is, but it's not a long time to have to keep resumes around like that.
So your risk is high.
The likelihood of it happening, that really depends on how you're set up. And if you aren't set up in a way that actually supports that minimal process that we talked about, you're probably more vulnerable to a hack like that.
So they all come together. All these things come together.
And the way you get ahead of it is to do the thinking at the beginning.
I'm all about don't make copies and do think,
[00:20:51] Speaker A: thinking campaign.
Okay.
[00:20:53] Speaker B: Revitalizing, reviving the thinking campaign from the late 90s.
[00:20:59] Speaker A: All right. Is that a wrap?
[00:21:02] Speaker B: I feel good about that. What do you think?
[00:21:05] Speaker A: We're 20 minutes into it, so I hope so.
[00:21:08] Speaker B: All right, well, next time I want to. I want to go back and I just want to circle back a little bit on the create data at closest to the source, as close to the source of the data as possible.
So the data that supports a business process should be under the control of the people in the business process. We talked about it kind of here and there through this, but I want to spend a little time on that specifically.
[00:21:31] Speaker A: Okay.
[00:21:32] Speaker B: Remember that for next time.
[00:21:34] Speaker A: If you have any questions, please send us an email to info trailblazer.com or sorry, info trailblazer us.com or find us on the web at www.trailblazer.us.com or check out the Learning Academy at trailblazer learningacademy.com no us in that one. Thank you for listening. Please tune into our next episode. If you like this one, please be a champion and share it with people in your social media network or like or subscribe. That would be great, too.
It's always helpful. We appreciate you, the listeners. Special thanks goes to Jason Blake created our music.
[00:22:17] Speaker B: Special thanks goes to Jason Blake, who created our music because you totally whispered there. So thanks, Jason, everybody. Have a great weekend or day.
Okay?
