Data Confessions: Case file secrets with Melanie and Nathan (Ep 005)

In this episode of the DATA SECRETS Podcast, Nathan Settembrini and Melanie Manning kick off a new series called “Data Confessions” – an inside look at the messy, thrilling, and sometimes hilarious adventures in real-world analytics consulting.

Nathan and Melanie pull back the curtain on their most intriguing current investigations, including:

• How to optimize business processes with tricky handoffs and unreliable data across multiple teams
• What it’s like to lead transformative analytics initiatives (and juggle both IT and business expectations on the ground)
• The surprising challenges in analyzing corporate real estate data: meeting room wars, “forks” as a KPI, and the reality of hybrid work
• Data horror stories, from multi-month cleanups (“Stop the Bleeding” and “Clean Up the Blood”) to “scream tests” on unused reports
• Moments when one metric changed everything (including the time a simple “cost per ton” dashboard blew the lid off a mining company’s profitability)

PLUS: The Lightning Round! Find out which vanity metrics, weird KPIs, and North Star measurements these pros love, hate, and use to drive results.

Nathan and Melanie also tackle big questions about operational process, the surprising accessibility of modern analytics, and what AI means for the future of dashboards and data transformations. If you’re a leader, analyst, or just data-curious, this episode is packed with practical lessons—and a few laughs—straight from the analytics trenches.

Chapters:
00:00 Opening Statement
01:30 Case File #1: Active Investigations
08:54 Case File #2: Evidence Gone Wrong
17:12 Case File #3: Breaking the Case
22:02 Lightning Round: Numbers on Trial
26:16 Closing Arguments

Comment with your own data confessions or questions—and don’t forget to like and subscribe so you never miss an episode!

Follow the DATA SECRETS Podcast

📬 Get episode recaps & bonus insights at allegroanalytics.com/podcast
🕺🏻 Connect with Nathan on LinkedIn
📺 Watch every episode on our YouTube Channel

Chapters

• 0:00 Opening Statement
• 1:30 Case File #1: Active Investigations
• 8:54 08:54 Case File #2: Evidence Gone Wrong
• 17:12 Case File #3: Breaking the Case
• 22:02 Lightning Round: Numbers on Trial
• 26:16 Closing Arguments

Transcript

Nathan Settembrini [00:00:00]:
Welcome back to the Data Secrets podcast where we uncover how business leaders crack the code, find truth in their data, and turn insight into action. Today's episode is something a little different. I have my co founder and business partner, Melanie Manning here with me. Melanie.

Melanie Manning [00:00:19]:
Hey.

Nathan Settembrini [00:00:20]:
And we're kicking off a new series that we're calling Data Confessions. So Melanie and I will sit down together, share some of our stories. Think of this as kind of a behind the scenes peek into some of the mysteries that we've solved in our work at Allegro analytics and throughout our careers. So we might tell some oldies that are goodies. So grab your flashlights. It's time for our first episode of Data Confessions. Your data has secrets, secrets that could change everything if you only knew where to look. Welcome to the Data Secrets podcast. All right, we're back. Melanie, how are you?

Melanie Manning [00:01:07]:
I'm great. Good to be with you, Nathan.

Nathan Settembrini [00:01:10]:
Yes, it's about time, guys, if you don't know, Melanie is the one who makes me look smart because she's so amazing and I love working with her and you would too, actually, if you were to hire us. Okay, so Data Confessions, we're going to do this in several different sections, case file sections, where case file number one is our active investigations. These are current projects that you're working on. We're going to need to redact client names to protect the innocent. And we just want to hear what's the business question, what's the data? And any surprising findings so far. So, Melanie, what are your active investigations or one of them?

Melanie Manning [00:02:03]:
There's a few. I think I'm going to focus on the one that to me is the most impactful to the business right now. Super exciting. The business question we're trying to answer is how to optimize a business process with multiple handoffs across multiple teams in the business. And it's really kind of the core function of what they sell. The current problem they're trying to solve is that the handoffs and the analysis and the data that they use is not reliable, not transitioning or tracked in a way that they need it to serve their clients. And so it's a really big project and it's part software development and it's part data visualization, telling their story in client facing reports. So it's a big challenge. It'll go into next year and tons of business discovery across the business. How to figure out how to optimize this process and get them to deliver the best service they can for their customers.

Nathan Settembrini [00:03:16]:
And so what, what is your role on this One, I mean, I know, but for the listeners.

Melanie Manning [00:03:22]:
Yeah, it's an interesting role and honestly it's my favorite role. So you know, you could think of it like a product manager like role but embedded with the business. So I'm really representing the business owner of the project and helping to tell their story across their team to our IT partners who are, you know, the architecting and developing this solution, this software solution and also helping to deliver requirements to them in such a way that we'll get the data that we need on the back end to tell their visual story in client facing reporting. So you know, this business owner of this project that I'm working with is brilliant. Knows everything about how the data, how the calculations should work, how the data should flow. Just is traveling constantly and very busy and very client facing and very critical in a lot of other capacities to that business. So she needs the kind of augmented support to make this project successful. And that's what I'm there to do, is to help her be successful and in doing so, help the client be successful.

Nathan Settembrini [00:04:36]:
Nice. Yeah. So it's more like management consulting, like traditional management consulting than more technical development. Like you're helping lead this initiative on behalf of the business.

Melanie Manning [00:04:49]:
Yeah, I'm part of a team that's helping lead IT for sure. That's awesome. There's a lot of other great folks on the front lines with me, business analyst who's doing a lot of technical deep dive inside their systems for requirements documentation. A really great project manager who's kind of spanning both sides of the fence, business and it. And I would say that my focus is really on the business side. So I guess representing that team but with the informed perspective of what the IT team needs from me. Right. Of the requirements that they're going to need and how they're going to need them specified so that they can build what we need them to build.

Nathan Settembrini [00:05:34]:
That's great.

Melanie Manning [00:05:36]:
So Nathan, same question to you. What projects are you working on and what are the business questions you're trying to solve?

Nathan Settembrini [00:05:44]:
Yeah. So we had a client in the corporate real estate and workplace space in big tech. And so you can imagine that big companies like Google and Apple and Microsoft, like they have a lot of employees and a lot of locations and they care a lot about the workplace where these people are effectively changing the world together, like where they come together to do that. And so you can imagine that at a company like that that there's a lot of data about the real estate both from what are leases, what are the critical dates that are Associated with those leases. How many seats do I have? How many people are assigned? Which people are assigned? Are they coming into the office? All these kinds of things. And this is something that I've been doing for the last decade really. It's kind of at this intersection of corporate real estate and data. Yeah. So for the last nine months we've been working with a client to build out an entire suite of tableau dashboards that and underlying data processes. So they have an in house data engineer and software engineer that we're partnering with to pull the data together. They're even using a lot of AI stuff, which is super cool. But our team is responsible for helping on the tableau development. And the most recent initiative is kind of focused on usage of the space. And so that's a lot of different data sources that we've kind of like pinned, built up over time where we have the capacity that's available. Not just desks, but meeting rooms and that type of thing. We have the people that are assigned to the office and. But with the business question they're trying to answer is, you know, at peak times it sure feels like we don't have any meeting rooms, which is a classic office real estate question. Um, and, and so there's, you know, you can book a meeting room, you can sense whether somebody is in the meeting room and then you have all this contextual information about how many, you know, what's the supply, demand and all this stuff. So all this comes together into, into one view to help the client understand, you know, for meeting rooms of this size in this building, on this floor from 10 to 12, we are 100% booked every Monday through Wednesday, that kind of thing. So it's been fascinating and yeah, it will be impactful because the things that it will drive or allow them to make better decisions on is the configuration of the space to support the work that's being done. So if the people are only coming in the office three out of five days, most of that work is going to be collaborative, which means the most of the space should probably be meeting rooms. And so that will help inform future build outs of the space.

Melanie Manning [00:08:50]:
Super cool, super relevant as we come back to the workplace.

Nathan Settembrini [00:08:54]:
Yeah, yeah, totally. Okay, so let's move into case file number two. This is evidence gone wrong. You know, in our normal interviews we refer to this as our data horror stories. But this could be a bad call that was made from data or some scary discovery or some mystery and like, you know, so what, what's your data horror story evidence gone wrong. That maybe you learned. What did you learn from it? How about that?

Melanie Manning [00:09:27]:
Okay, so a data horror story. To me, like in one way it was super painful, multi month, big effort, long hours kind of story. But in another way it was thrilling because in the end we solved it and it was super impactful to the business and really changed the way they looked at how they sold the business. Problem was, are we pricing our products well? And this particular company, they had a super complex service delivery model. Tons of components to what they sold. Each component had a price associated with it. But then they would bundle them and then they would discount, apply discounts based on what got bundled in. And it got very convoluted as to, okay, well, what are we making on this client? Really? Right. Because tons of teams across the business were having to deliver on this, this product. None of those teams were billing their hours against the project itself. So we, they had no visibility to the profitability of an individual client. But yet the sales team was being incented to close all the deals regardless of the discount. And the discount wasn't even that clear because of how many different products came into the mix. And so the reason that it was a horror story was because we came into this problem with like starting at zero. There were, there was no project system being used and there was no discipline across a team of hundreds of people to collect data that we needed for this problem. It was so big that the profitability of the company was at risk. And so we came in and did some discovery and kind of the outcome of that discovery was a two part plan and to just tell you how like big and painful the problem was to them, like phase one, we called it stop the bleeding. And phase two, we called it clean up the blood.

Nathan Settembrini [00:11:39]:
Oh my gosh. So officially those were the official project names.

Melanie Manning [00:11:46]:
Yeah, I mean that was phase one.

Nathan Settembrini [00:11:47]:
Phase two, that's hilarious.

Melanie Manning [00:11:49]:
In that order, because we were like, okay, we like, we gotta, we gotta first like stop letting the problem get worse, right? So phase one, we implemented a project system. But not only that, we had to implement a whole operational process to use it, to require it, to require certain data fields at certain points in the project, all that. It was a huge operational transformation to implement that system. Began collecting good data, right? So pretty pretend, like on day 10, we start collecting good data. It actually was probably month 10, maybe, maybe six to 10 months. We got to that point of implementing that system and then, okay, good data is coming in, but we need three years of history too. So that was phase two, Clean up the blood. Go back to three years of history on a company that was near a billion dollars in size. Okay. And clean all that data and get it to. And not in an after system, but in the source system so that it would flow through to reporting so that the business could have visibility to how profitable the customers were and the outcome is positive. So in that sense, it's not a horror story, but in the sense that it was a big hairy project that took a long time to accomplish, gave all of us who worked on it quite a lot of perspective and really changed the way they modeled deals going forward. So worth the effort, worth the heavy lift for sure.

Nathan Settembrini [00:13:28]:
Nice.

Melanie Manning [00:13:29]:
How about you?

Nathan Settembrini [00:13:32]:
Yeah, mine. So we'll go back to 2013. 13. I was sitting in my cubicle at Equifax. This is, you know, way pre Covid. So, you know, we couldn't even work from home if we tried. I was supporting our inside sales department. We had about 50 people on the phone. And I was responsible for, it's kind of like sales enablement. And so we did some technology implementations. But then also I was responsible for reporting and mostly pulling data from different systems and smashing it together in Excel, as one does when you don't have the proper tools. So there was this one report, I don't even remember what it was, but I inherited it from somebody else. And every Monday I would have to put it together and send it out to this distribution list. And every single time I sent it out, crickets. I never heard a single thing back from anybody ever on this report. So I started working there in 2012, I believe. And I've been doing this every Monday for weeks and weeks and weeks, months and months. And one day, one Monday, I came in, I just had a wild hair, and I was like, you know what? Today we're gonna do a scream test. I didn't call it that. I didn't know to call it that at the time. But if you don't know, a scream test is when you either take a tool away or you just, you know, stop providing something and see if anybody screams. And so I did that. I basically, Monday, 10am came around. I was like, you know what? I'm gonna see if anybody cares about this report. And, you know, lunchtime rolls around, I'm like, okay, no, nobody's asking. Head off to lunch. And I'm like thinking about it the whole time and then come back after lunch. I'm like, surely there's going to be an email waiting for me and not end of day, nothing end of the week. Nothing. Weeks go by, weeks go, even months. And then eventually one person, one sales leader sends me an email that was like a reply all to one. Like the last one that I had sent months before says, hey, do you have this report? Are you sending this out anymore? Because I need this one number off of this one cell. And so I was like, oh, well, I can get you that one number for that one. And yeah, that. That was the. Basically the. The horror that happens in a lot of big corporations is that there's all this reporting that doesn't get used. You know, we like to call it shelfware because it lives on the shelf. Nobody wants to build a dashboard or report that doesn't get used. And so I would encourage all the analysts out there, if you feel like no one's using one of your reports, just stop sending it or just archive it. See, you know, do the old screen test. So there you go.

Melanie Manning [00:16:55]:
Wise words.

Nathan Settembrini [00:16:58]:
But, yeah, this is not legal advice. Do not sue us.

Melanie Manning [00:17:04]:
Just do it at your own risk.

Nathan Settembrini [00:17:07]:
You know, if you feel like you need to get your boss's blessing to test it out, do that. All right, case file number three, Breaking the case. So these are stories where the client data revealed something unexpected. You know, what was the big realization? How did it change their path? Or like, what was the business impact for you?

Melanie Manning [00:17:30]:
Okay, so this example is more recent. I'm working with a client who has a legacy data transformation process in place, and they're using Snowflake, and in the process of that data being pulled into Snowflake, it's being transformed and running through another system to run calcs on it. And by the time it makes it to the end user, it's quite altered from its original state. And this customer, they came to me and said, I think that there are things in the original file that I actually need. And I think some of those transformations are actually doing things I don't want them to do. So can you help me to get my hands on that unaltered data before it goes through all the transformations? What's it look like? So we went and did a little work. We worked with their data engineering team to get access to some places in their data warehouse that are kind of locked up. And we exposed that to the client and gave them a little bit of training on how to write SQL queries so that they could, on the fly, dig into whatever their hearts desired. And they did. They learned that a lot of the transformations that were going on were not even necessary because the information they needed was in the source Data and that. That. Yeah. So it has like changed their approach to how they're planning to use this data and possibly even abandon the transformations.

Nathan Settembrini [00:19:16]:
Wow, nice. Yes. Mine is, it was a mining company from a couple years back. So you can imagine mining your doing a lot of work to dig up a bunch of stuff and you find good stuff inside. But how much did it cost you to produce one ton of the good stuff? They could not tell you what their cost per ton was. And so we did a discovery with them, teased it out, came up with the calculation with the team. Then we built out the data prep to calculate this metric along with several other metrics and put it on a dashboard for their executive team. They had never, not only had they never been able to calculate this metric, they'd never seen it, much less seen it trended over time. So we, you know, we hooked into the source systems and it's fascinating. So, you know, sometimes when you build a dashboard, you will, you will build it based on kind of on flat files, some historical snapshot that you pulled maybe at the beginning of the project. And meanwhile the data engineering team is building the actual pipelines, building the actual queries to combine the data together. And then at some point you do this hot swap of the data source. And so we did that for the first time and we saw cost per ton was just going up, up, up, profit was going down, down, down. And we, so we, you know, it didn't take much to read the, the tea leaves there. You know, it doesn't take a genius to see what's happening. And we got a little scared because we were like, you know, we just spent all this time, you know, we've got this huge invoice that, that's outstanding for this client. And then, you know, they came to us and they said, hey, our business isn't doing so good. We're like, sorry, I'm laughing now, but it was like, oh, my gosh, we're so sorry. They're like, we really do want to pay you. Could we pay you over time? Could you put us on a payment plan or something as we figure out how to, you know, fix our operations or what have you? So, uh, they did end up paying in full over time, but it was just so, so amazing how the data. Yeah. Told the story. And then that story came to life for us. It was kind of the. Not quite a canary in the coal mine because it was so obvious, but.

Melanie Manning [00:21:55]:
Anyways, also a story of being flexible and customer friendly when you need to be.

Nathan Settembrini [00:22:01]:
Yeah, totally. Let's transition into our lightning round. We're calling this numbers on trial. So which metrics are guilty or innocent? So we're going to go through a few different, like, categories of KPIs. Just three and yeah, as fast as we can rattle them off. So let's start with vanity metrics that you hate or secretly love.

Melanie Manning [00:22:30]:
I'll say the annualized monthly revenue just based on one month.

Nathan Settembrini [00:22:37]:
Oh, interesting.

Melanie Manning [00:22:38]:
I think it doesn't tell a complete story and you can infer a lot from it that doesn't prove out.

Nathan Settembrini [00:22:47]:
Yeah, that's interesting. Yeah, because you tend to focus on those during the good months. You're like, oh, we had a peak. Like, what's our annualized revenue based off that peak?

Melanie Manning [00:22:57]:
Based off that one month?

Nathan Settembrini [00:22:58]:
Yeah. For me it's LinkedIn impressions on posts. The algorithm has been beating me up real bad lately and it hurts my feelings, but I have to remind myself that just gotta keep putting out good stuff and it'll be fine. It'll be fine.

Melanie Manning [00:23:20]:
Just keep doing the right thing.

Nathan Settembrini [00:23:22]:
Yep. What's the weirdest KPI that you've ever seen?

Melanie Manning [00:23:27]:
Okay, it's just weird in name, but it's actually a commonly used KPI. But Junk Per ton was a KPI we used at a client. It was just funny that it was junk per ton, but it's an actual term in manufacturing, so.

Nathan Settembrini [00:23:44]:
That'S awesome.

Melanie Manning [00:23:45]:
How about you?

Nathan Settembrini [00:23:48]:
In the corporate real estate space? There's some funny ones. So forks is, you know, you can imagine a culinary operation at a large tech company. How do you count number of meals? You know, you could call it meals, but they call it forks. I don't know, I think it was from, you know, let's just count the number of forks that we have to keep replacing in the course of a meal. Another one is carpet per square foot, which Jeff Gegnon told me about, which is, you know, if you think about all of the space that a tenant occupies, what it's kind of like, what percentage of that is carpet revenue per square foot is a. Is an interesting. It's not a weird one, but it was, it was kind of unique. If you think about a company that has all this space to produce some widgets, you know, what is the revenue generated per square foot? It's kind of interesting, but anyways, I still like forks, but why not spoons? I guess every. Everyone uses a fork for the most part. I don't know. Except for on sushi day. What do you do then? Okay, what about the third one? Last one is your north Star.

Melanie Manning [00:25:11]:
Okay, so probably, probably two. I think the bookings gap to budget is super important to me. If you forecast out the whole year and based on what's under contract, what, what you know, that you've sold and just need to deliver on, you know, take budget minus that for the year and figure out, okay, well, this is what we need to sell to hit our budget goal. That one's, that one's a great one to me. I think it tells a good story. Another one is days. Cash on hand course. Right. Cash is king.

Nathan Settembrini [00:25:48]:
Love that one. For me, lately it's been sales meetings. So how many conversations are we having every week with new prospective clients? That's easy one to track.

Melanie Manning [00:26:02]:
Also. Great way to spend the week.

Nathan Settembrini [00:26:04]:
Yeah, yeah. I mean, I love, love talking to clients. Which by the way, if you'd like to reach out, please do. Nathan Allegra analytics.com. okay, we're now entering. Gonna land the plane. Entering our closing argument section. You know, we don't have to hit all these in great detail. There's only a couple. But is there one thing that you wish leaders knew about their data or understood more about their data?

Melanie Manning [00:26:35]:
I wish that leaders understood the full impact of operational process on their data and designed their processes to of course be efficient and effective for their clients and also to configure their systems to enable good data collection to help them measure success. I just think we run into that a lot where data collection is just not available in a really important scenario that the client really needs to know the answer to a question and they just don't have their operational processes aligned to deliver that answer. So it just becomes a longer tail project to get that alignment. But I wish that that was kind of factored in on the front end as processes are implemented with our clients or with any company. How about you?

Nathan Settembrini [00:27:33]:
I would say I wish leaders understood that getting started with analytics is actually way less expensive than it used to be. Borderline cheap, especially if you take a scrappy perspective on. Let's focus on one business impact. Let's pull the data through and move the needle there and then kind of expand. You don't have to start by building a Lamborghini. You can start by building a bike to get you somewhere, which is way better than mashing Excel spreadsheets together like I was doing in 2013. And then later we can upgrade that bike to a Subaru, you know, and then maybe someday build out the Lamborghini. But building out data infrastructure is not a million dollar investment like it used to be. And what about AI the future? Do you have any kind of reflections or thoughts there?

Melanie Manning [00:28:47]:
So in our space I can see what we developed today and what we've historically developed for the last 10 years has been a dashboard that delivers the whole answer. Right. A visual story, leveraging data that is kind of fixed in nature. Right. We always customize it to our clients needs and give them the ability to filter for their relevant data set. We design the charts that are on the dashboard and the data underlying to tell their unique story that they tell us they need to hear. But I think as AI gets embedded in these products more and more, I think that what we're going to see is maybe a higher level, more generalized dashboard, or maybe not even higher level, but more specific dashboard feature with a ability to ask questions in natural language right alongside of it.

Nathan Settembrini [00:29:50]:
Yep.

Melanie Manning [00:29:50]:
And those of us who are in this space will be implementing the business context to those metrics and that data in the semantic layer so that it can feed those responses in natural language. So it, you know, our, our work is shifting for sure, but I think that it'll benefit the end users and I mean even the ones, the non analyst end users who just want to ask a question and get an answer. I think it's really exciting.

Nathan Settembrini [00:30:22]:
Yeah, it really is. And so I think on the other hand, one thing that I've just noticed is that a lot of companies are kind of pumping the brakes on any sort of investment in anything around data because they're just kind of assuming that AI is going to solve all their problems. And so I would just encourage folks that in order for AI to do anything, the data's got to be together and there has to be some level of understanding. They call it the semantic layer, which just means a level of understanding or layer of understanding about the data, kind of interpreting what those things mean so that the AI can sales up good, cost per ton up bad. And so in order to get there traditional bi, bringing the data into a data warehouse and adding some of that context, that's really a great stepping stone towards being able to leverage these AI tools in the very near future. And I agree that our jobs as analysts and that sort of thing, we used to be dashboard builders, tableau experts. It's going to shift and that's okay, I think, as long as we keep learning and keep, keep growing and keep trying to serve our clients well. So Melanie, this has been super, super fun, super awesome. I would encourage our listeners to connect with us. You know, we'd love to have a conversation with you, you comment if you're on YouTube, comment if you're listening on a podcast. We also publish these things to YouTube and we'd love to respond to your comments in YouTube. And don't forget to like and subscribe because that helps other people like you hear about data secrets. Yeah. So until next time, keep searching. The answers are out there. That's the pod. It.