213 - Unleashing the Power of Data: A Journey to Transforming Software Development with Allstacks' Hersh Tapadia

Welcome to Listening with Leaders. I'm Doug Noll, lawyer turned peacemaker. I teach executive leaders how to listen to emotions rather than words so that they can become the leaders everyone wants to follow. And I teach those same leaders how to be authentically present, available, and connected to their families despite being insanely busy. I have learned that we are 98% emotional and only 2% rational. Learning how to listen to emotions is, in my experience, the foundational skill of life. Stick around to the end of the show and I'll reveal how you can be on our next guest in 15 to 20 minutes. So let's get started. Hersh Tivadia, welcome to Listening with Leaders. You are a serial tech entrepreneur, but right now you are the CEO and founder of a company called Allstacks. Which can be found at allstacks.com. Thanks for joining me.

Thanks for having me.

You have a kind of a long list of companies that you've been involved with. Tell us a little bit about your backstory.

Yeah, absolutely. So I, short version of long story is essentially I started my career at IBM and the, the hardest part about being at IBM in 100,000 engineer organization is that I couldn't find any connection to the value of the work that we were doing. And I really struggled with, does my work matter? Can I progress here in my career? Can I get to anything interesting? And IBM had this tool at the time called Blue Pages, and Blue Pages was like an internal LinkedIn, but it had this interesting piece where it would show you the number of layers of management between you and the CEO.

Oh, wow.

I was like software engineer 1, and I'm, you know, going scrolling, scrolling, scrolling, scrolling, scrolling, scrolling through 2 pages of people. And I said, gosh, like there's no way, there's nothing that is encouraging about this situation. And then, so I left there and I was trying to figure out what to do. And I ran into a neighbor and my neighbor had a startup. And it was in the signal processing defense area. And I joined and it was a fascinating picture of hypergrowth. And so before I knew it, this 4-person group had turned into a 20-person group. It turned into a 100-person group, turned into a 400-person group over just a very short couple years. And there was even this physical component of it where we had, we were in a strip mall. And we had one bay of the strip mall and then we had two and then we had four. And then next thing I knew they were building a four-story building and then they were building four buildings and it was like very visibly growing and it was super fun, very addictive. And that company did great. The founders made a ton of money, the employees made a ton of money, and I said, this is the thing, this is what I want to do. At the same time, when I was going through school, we got the opportunity to start a company instead of doing a senior project in electrical engineering at NC State. And so what I did was we stumbled onto this concept of this idea that when you're diagnosing tuberculosis, similar to COVID actually, The speed to diagnosis is the biggest problem. And so you're contagious right away when you have an active tuberculosis infection. So the faster you catch it, the less it spreads. The way you catch it is you stick a stick down somebody's throat until they cough up their lung fluid, stick it on the slide, and you look at it and you hope you can look at enough of the slide to figure out if somebody's infected or not. It's notoriously unreliable. So the big plan was you do it 3 times over 3 weeks and hopefully disease had progressed enough by that point that it was very obvious that you were sick. Meanwhile, you're like in a bus in India coughing and the whole bus is getting sick. Huge problem. So we had this idea that we would use computer vision on a microscope camera to analyze 1,000 times the amount of sample in one image that a manual operator could do in 3 weeks and thus detect it early. And so we were building that product. It was super interesting. The world economy collapsed in 2009, but we figured out a way to get it sold and joined an organization doing supply chain security for the counterfeit drugs market. How do we secure the drug supply chain so that counterfeit or what's called diluted drugs, right? Cut pills, things like that don't end up in the hands of patients that need those treatments. And so we got to use that computer vision technology and I was running engineering, I was running product there and we used the technology to scan the pills and try to track these drugs through supply chain and tell you where is it a real drug? Is it a fake drug? Is it in the wrong place? This drug was supposed to be in Greece and it ended up in the US or whatever. And eventually started to expand that concept to other things. And in that, I had this interesting experience where I was explaining to a set of scientists how software development worked. And we were, because we were building software and it was challenging, not because everyone's smart in these fields. There's a language barrier. There's a lack of familiarity. Vocabulary is challenging, not necessarily like English, but like Literally, like, when you say test, you mean something different than when I say test. Right. And so that was challenging. And then we didn't have data. We're building these data products. We didn't have any data that would allow us to solve these problems. And ultimately, what I felt was that there's a problem of communication here. And what the data creates is this baseline that then helps two parties communicate. Set from the baseline to where they want to get to. And if we could solve that problem, it would be really interesting because how could it be that software development is the only part of the business that's not quantified, unlike everything else— marketing, sales, even like customer success and all that? And so we're like, okay, there's something interesting there. And we left. I was like in my mid to late 20s, and I said, I don't want to spend the rest of my 20s here. I want to to another thing.

I need to pause, I need to pause one sec. But you had, you now were looking at how do we track software development? I think that's what you were—

I was reflecting on my experience of trying to explain the outcomes we were creating and the challenges we were having in developing software, but I didn't have a good way to do it. But I thought it was a solvable problem because we are as a, as an organization, like as a group of an industry, software engineers are generating tons and tons of data, but the data is not being used to understand process and commitments and capabilities and forecasting the same way we're using that data in other places, in sales and marketing, etc. So I think there's a, there's an opportunity here. And Jeremy, who's my CTO, him and I set out and we said, we're going to solve this problem. And the first thing we did was we went interviewed hundreds of engineering leaders. We found that everybody was struggling with this challenge and there's a real product here. And so we started building it.

And that's what led to Allstacks.

And that's what led to Allstacks. So we harvest the data from these underlying tools that developers use. So tools like GitHub, from Jira, the Microsoft equivalent, Atlassian has their suite, Microsoft has their suite, and then there's like GitLab and a bunch of other tools. And we harvest the data from those tools. Just things like people going about their day generate a trail of activity. And when we look at that activity on a broad, almost epidemiological level, we see trends that allow us to understand, are we on track or off track? How do things typically flow and where are we deviating from those flows? And then if we can make that— those processes as efficient as possible, then you get the opportunity to say, okay, we have this pipeline of building software. It's as effective as could be. Now let's make sure it's pointed at the right things, the most important things. And then once we know that our really effective pipeline is pointed at the really highest value outcomes, then the business really starts to prosper. And that's—

That strikes me as being a mission-critical tool.

I think it is, especially at scale, right? It's an interesting thing where at the end of the day, it's super important in the right— when you're in the right place at the right hierarchy of needs. We talk about this a lot with our customers and our prospects. We said, look, if you're at a place where you're not sure that you can, you can send some software out the door without bringing down your whole system, you probably don't need to worry about optimizing flow. You need to just get the foundation repaired, right, of your house. But once you do, then you can go out and solve these problems, right? I had a mentor that told me about— he said, he said he would tell these parables. Super good story. Basically, he'd say, okay, you're— let's say you own a beach house. And you're sitting there in that beach house and you wake up in the middle of the night and you're all stressed out because you just had this terrible dream. Climate change is happening. Sea levels are rising. The beach house is about to fall in the water. When? I don't know. Sometime in the future, but, but it's stressing us. Okay. I'm gonna calm down. I'm gonna go downstairs, get a glass of water. And as you're getting out of bed, you realize the house is on fire. And you gotta run to the fire extinguisher, which is in the kitchen. And so as you're like running down to get the fire extinguisher, there's a guy on the stairs with a knife and he's trying to stab you. So in that moment you say, okay, what problem do I need to solve? You need to deal with the guy with the knife before you can solve the fire, before you can worry about climate change. And so one of the things that we help our customers with is which problem are you solving right now?

Priority. Mm-hmm.

How do you prioritize? They're all important, but on different horizons. Right.

Different urgencies. So what Allstack does is it allows program managers and engineering supervisors to take all this, harvest all this data, as you said, the, the data leftover from activity by software developers, programmers, coders, and then they get a big, they get a high-level picture of what's going on and allows them to make better management decisions.

Yeah. And there's critically levels of engagement to the software. So think about it as if you're a CTO, right? Exactly. What you're, what we're helping you do is understand where are we investing? Is it aligned to the things that matter to the business? And when they're not, is it a systemic issue? Or is it an acute issue? In the middle management level, you're saying, I have made commitments to other stakeholders, CS, sales, marketing, etc. Are we going to deliver on those commitments? Will the teams that are tasked with those commitments actually be able to do it on time, on budget, or is that changing? And if it's changing, let me get early warnings. So I can go and talk to the people I need to talk to. And then at the line level, at the first line manager level, what we're saying is you may not get to choose what you get to work on, but you can make your teams as effective as possible against those commitments that, that have been made. And every day, we'll help you identify the key abnormalities, the anomalies, that aren't in line with the processes that you've brought your team up to, that keep you from delivering in the way that you've gotten to. And so it's like these small incremental habitual behaviors that we build inside of our customers. They help our customers essentially co— develop a compound interest strategy to success. Because we do a little bit every day to stay on track and get a little bit better every day. And that builds throughout the whole organization. And so that's what I was saying earlier about if we can make our pipe as effective as possible and then we can point our pipe to the best things to work on, the business is going to prosper.

And how's it being accepted in the marketplace?

Oh, it's great. It's so interesting. Because this is not a new idea. This is— there's a textbook on the bookshelf here in my office from the early '90s. It was a book around software development metrics written by a team out of AT&T. And if you read the book, you would think it's a modern piece of work. You had to swap out the screenshots of old Windows 95. Windows 3.1 type UI, but take those out and you'll say, oh wow, like we've been talking about this for a really long time. The difference is twofold. One is that we have means to acquire the data without being disruptive to the people doing the work. That's super important. We cannot go and ask a developer, take time out of your already busy day to update a bunch of statuses so that we can harvest data.

Right.

Only do the work that you have to do anyways to work, collaborate within your team. We'll take care of the rest. The second thing that's really important, especially right now, is that money is not free anymore. So COVID was an interesting anomaly in that the ability to hire ability to make investments with high expected value, but very high risk. That was very high, right? Because we could take moonshots because the marginal cost of doing it was very low. We're not in that world anymore. So before we said, I don't know if we're being as productive as possible. And if you think about the story with the knife and the fire and everything, my question, are we being as productive as possible? I'm not sure. We should hire more people to be more productive. Why did the conclusion become hire more people instead of improve process? We have this parallel problem was we can't hire people anyways because the economy's on like doing so well and people are scarce and employment numbers are through the roof and we're competing on dollars and we can't afford to lose who we have and we gotta grab everyone we can get in that world. The climate change was the, the productivity problem, right? And the guy with the knife was, can I even hire someone?

I've got manpower needs right now, or I need to be defensive about this because if I don't hire people right now when I do need them, they're not going to be available.

And so what that created was a situation where we didn't worry about efficiency and productivity very much. Some people did, right? Maybe at a macro level. But now we're not in that economy and there's no more recs. You're not getting more staff, but we still have to deliver. We still have to do what we can do. We still have to keep the business running. So what do we do? We have to get efficient. We have to get productive. We have to get focused. And so whereas before it might have been okay to work on the moonshot because the ROI, the potential ROI working on that moonshot was so high, that were willing to take the risk even if 80% of them failed. Now we're saying no, we got to focus on the core business. And every minute that's spent not working on the core business is risk. And so focus matters a lot. And there is a critical need at the management level, the executive level in particular, to say But are we sure we're focused on the right things? Can you show me that you're focused on the right things? For the sales leaders, they're saying, look, let's look at the pipeline. Are the deals in the pipeline the types of ICP customers we need to sell? For engineering, it's all the— are the features that you're working on aligned to features that would help push the top 3 OKRs of the business? We have a goal to reduce our COGS and the features we're building aren't going to help reduce COGS, then why are we building them? Right? And that the onus is on the departmental leaders to come to the table with the data to prove that they're doing what they're doing. And that has been a part of the trend that is driving our industry, not just Allstacks, but our industry as a whole. Wow. Right.

So what's your role in the company? You're the CEO and founder, so you're giving overall direction, the typical stuff. As you're working through this, what is it that gets you up in the morning and gets you excited to go to work?

Yeah, I'll say this, like one, I love the category. I love what I do and I love the people I work with. So it's very easy to get up. It's very easy to get up and go to work with this team of people and say, yeah, I want to make all of us successful. It's not even a question. But what gets me excited specifically goes back to this initial insight that we had years ago, which is at the end of the day, these are not problems of data or efficiency or productivity. They're problems of communication.

People.

They're people-to-people interactions. And so I think that if what we're able to do is help people move the ball forward because we've reduced the friction from poor communication, where it's not even intentionally poor communication, it's literally we're using different words, we're on the— not on the same page, or we're not able to establish a shared sense of truth because we all have our own perspectives. So this thing is going slow, this thing is going fast. Well, is it going slow or is it going fast? I don't know. What context are we speaking from? Is it that it's going faster but it's still slow? Both of those things can be true.

Context.

Look at the data and see what it looks like. And then we can establish, given this data, are we happy with it or are we not? If we want it to change in what direction and what derivative direction and so on and so forth. And that's a fascinating problem to solve.

It's so interesting. We do the same thing. I work with eliminating the friction in communication, brain to brain. And I teach people how to listen to emotions because we are emotional beings. We're not rational. There's no such thing as rationality. We can use reason and critical thinking skills and all the stuff that you're highly skilled in. And I, from my background in law as a lawyer, have obviously strong analytical skills. But it's when you— what do you do? What's the technology you can use to reduce the friction in communication? And it's so interesting to me because what you're talking about is exactly what I do. Only not at the technical, not at the, in the way that you do it. But you have articulated, I think, one of the fundamental problems in our society, and that is the friction of communication. And that's why there's so much polarization, for example.

Exactly. Can we understand where we're coming to the conclusion, how we're arriving at the conclusion? Because if we can understand how we're arriving at the conclusion, then we can understand the merits or the relative merits of the conclusion. Then we can take the opportunity to say, was your path there? Did the path itself have merit? We can start. It's really interesting. One of the things we talk about in the product is we, we say, here's a forecast of when we think something's going to be delivered. We don't just give you a date. We give you a date, and then we give you a change, and then we tell you over what time period did it change. So take two scenarios. One is something's on time and something's late. You say, well, if we put this on a RAG status report, what would you say? This one's green, this one's red. Okay, cool. The one that's on time, Every day is getting a week later. And in 3 weeks it's going to be late. It's due in 6 weeks, but in 3 weeks it's going to be— start to get forecast late.

You can see that in the data.

And you can see that in the data. The one that's red is late already, but the date hasn't changed in 3 weeks, and it's late but it's stable. So if at that point somebody came to you and said, hey, there's this green one, there's red one, what do you bet on? Do you bet on the green one that's turning red or do you bet on the red one that's staying red? It's a very different conversation, right? Hey, yeah, we're late already, but we're not gonna miss again and we feel confident this is gonna happen. This one, we have no idea. Like we happen to be early. Because we haven't run out of time yet. Totally different level of confidence. And it changes the conversation, right? So we, we try to give our customers, give our users that level of context to the information we convey so that they can convey us a narrative that isn't static, right?

Do you teach them how to use the tools to communicate? Here's how, here's the data. And this is how you got to use it, and this is the way you got to think about it. And this is the— in framing it, the kind of the high level that you're talking about here.

Oh, absolutely. One of the things that we're known for is that specifically. Our customers look to us as their thought partners. Our customers work with us to tell a story and to develop a narrative. The data is the vehicle for that.

I get that.

That one of my proudest moments for us as a company was when one of our customers had our CS person, their representative, come out to their annual engineering kickoff. And they flew her out there. And they said, will you actually present the story? To our team and set the tone for how we're gonna run with data this year. And then they invited her out to do that because they said, you've been such a strong partner for us that we actually want you to represent yourself here.

Wow. So how do you train your people to be able to do this, to be able to create the narrative, this and the story out of the data And teach the customers how, how to use the story themselves and how to put context around all of this.

Yeah, it's very interesting because there's several layers to how you do it. So first we have to hire people that are willing to engage in storytelling, right? Part of that is the, the person that you hire has to have a proclivity towards storytelling. That doesn't mean they're trained on everything, right? But you can see, I can lead this person to this place. And so we're looking for people like that. The next thing is they have to be able to engender trust. Ultimately, these things are built on trust, right? And so we have to be able to build a relationship with our customer. Part of it is the product has to work. Part of it is the product has to show valid data. Problem with a lot of data products is that they come in— there's this phrase, garbage in, garbage out.

I remember, I started coding in 1969.

Exactly, right? It's like I said, there's nothing new in the industry, right? But why does that phrase exist? Because especially with data products and dashboarding products, they'll come in and they'll say, my inputs are bad, so why— how can I trust anything here? So it starts with a product, right? We have to tell a story with data that is good. We have to help the customer understand that we've designed a product that finds the good in the data, even if all of the data isn't good. But we still find the good in the data so we can represent what you have, right? We say we'll meet you where you are. And then from there, we'll make these incremental changes, but only the ones that benefit you to get, continue to have better data, right? So it's a, it's an incremental trust building process. And then we tie to, you just ask an open-ended question, right? What are you being asked for? What are you having to explain? You're going to this ceremony, you're going to this product owner sync next week. What are you gonna tell them and what do you need to tell them? What do you want to tell them?

What's the outcome you're looking for?

Yeah, what do you want to tell them and what do you need to tell them?

Right.

Right, and sometimes the delta between the two, that's where you get to develop the narrative. That's where you get to tell them the story.

So that strikes me then that listening is one of the most critical tools that your people have when they're working with customers.

Oh, absolutely. Absolutely.

And then knowing how to listen, what to listen for, what's not being said, what is being said, how are people expressing their emotions around all of this. And it becomes a, a really critical skill for people to develop.

Yeah. People rarely come to a product like ours, and then there's product like ours for multiple categories, right? But you rarely come to a product like ours and ask a question about what that's actually about what you're seeing. There's some other question behind it. They say, if they're looking at a graph and they say, why is that line going down?

When people ask about what they see, oftentimes what they're asking for is, or what they're asking about is something that they don't, that they don't, don't either, they don't see it or they're seeing something completely different and they're not able to articulate well where they're going.

Or there's some emotion that they're feeling when they see this that's a result of something else.

Yes.

And they're trying to unpack that. And if you can get to the root—

Right, anxiety, fear, concern, maybe frustration.

If you looked at a graph and you said, why is that line going down? Right? And their expectation is the line's going up or it's flat. They're not asking about why is the line going down. They're not asking about like, why is the data lower this week than it was last week? What they're asking is, are we going to hit our commitment? What am I going to tell my boss when I have to tell them that this is going to be late? Right.

Anxiety, a lot of anxiety.

Can I recover this? Can I solve this?

Worrying.

And what the data, what data products like ours, what Allstacks allows you to do is get ahead of it. I'm not surprised, right? If I can give you, if I can get that feed consistently with regularity over time, and I'm not telling you a snapshot, I'm not telling you January 1st, we said we're going to ship this feature. And at the end of the quarter of March 30th, March 31st, we didn't do it. It happened in between. So that every week we're saying it's getting a little bit later, it's getting a little bit later, then uncertainty here, it's becoming harder to work in this code base. This part of the product's very old and brittle. And what we're finding is that every time we change something here, it's then creating 3 more things we have to fix. And so the scope is ballooning. Because this is an old piece of bank software that nobody's touched in 25 years. Right. I mean, so then what? Actually, probably what is we need to start from scratch on this piece, not try to keep band-aiding what we have. And if you're doing that January 1st, you might be saying, how do we make that trade-off? Right. And what is the ROI? Is it really worth doing? It seems like the easier path is just to mandate it. But if you have the data on January 31st, you might say, we need to abandon this path and spend the next 2 months doing it from first principles. If you don't have the data, you might say on March 31st, I wish we had abandoned the path 2 months ago.

I teach a course in decision-making at Pepperdine, a graduate course on decision-making under uncertainty and conflict. And one of the things that I teach is how important information is in the decision-making process. You can have too much, but most of the time you don't have enough.

Mm-hmm.

And so in this context, when you're trying to decide what, what are we gonna do, the data that Allstacks provides is critical to a good decision-making process. And without that data, it's pure guesswork.

Yep.

You're pooping a coin. Do we stay with the old stuff or do we go new? And what you provide—

A lot of the people I talked to before we started the company told me their best tool was walking around the floor back when everyone was in office and trying to figure out just from sampling where everything's rare. And they said, you know, Sneakernet.

They say, "SneakerNet's my best tool." And, and of course when the pandemic came and we all went online and everybody went remote, that became impossible.

Right.

And now there's no way of really checking except to see whatever work a software developer coding might do. You, I guess you could look at it on a weekly basis, but that's a lot of data to absorb if you're a program manager, project manager.

Absolutely.

Wow. We're coming to the end of all of this. This has been absolutely fascinating. Conversation. The key that I hear is that your project is not about data. It's about how to understand the data and have intelligent conversations based on what the data shows. And that's really more important. That's really the more important thing for your customers to be doing rather than just thinking about the data. And it's your tools that give them a foundation upon which to have these critical conversations that they otherwise couldn't have or wouldn't have effectively. Is that a fair summary?

Absolutely. I'll leave you with this story I tell the students when I give an annual talk at an entrepreneurship program at NC State. And one of the stories I often tell is that I'll ask them, they're all engineers, and I'll say, I have a hammer. Why did I pick up this hammer? And the answer is usually you gotta— you're trying to hammer in a nail or something, right? So no, I'm not trying to hammer in a nail. Yes, literally, I might be trying to hammer in a nail, but I'm not picking up the hammer with the goal of hammering in a nail. I'm picking up the hammer because I want to hang a picture, preserve a memory, be reminded of something when I come down the stairs every day. And that's the point. The tool, the data, all that is the nail and the hammer.

Right.

But the goal is, what do I see? Oh, I see that picture I took of that village in Italy that I was at. Right. And I want to remember that every morning when I come downstairs.

I, I think that's a great perspective. Amazing conversation. Thank you so much, Hersh. This has been really incredible. I've learned a lot.

I enjoyed it a lot.

Thank you so much.

Thank you for having me. Absolutely.

Doug Noll here. Thank you so much for listening to Listening with Leaders. If you are a successful executive leader who would like to be on this program, please visit podcast.dougnoll.com.

DougNoll.com/podcast.

If you got something out of this interview, would you please share this episode on social media? Just do a quick screenshot with your phone and text it to a friend or post it on the socials. If you know someone that would be a great guest, tag them on the social media to let them know about the show and include the hashtag #ListeningWithLeaders. I love seeing your posts and guest suggestions. We are regularly putting out new episodes and content to make sure you don't miss any episodes. Go ahead and subscribe. Your thumbs up ratings and reviews go a long way to help promote the show and mean a lot to me and my team. Want to know more? Go to my website, DougNoll.com, or follow me on LinkedIn, Facebook, and Instagram. That's @DouglasEnoll. Thanks for listening, and we'll see you on the next show.