Lessons from Corporate Innovators

Keith Berry Executive Director — Moody’s Analytics Accelerator

Play episode

Episode 19: Keith Berry Executive Director — Moody’s Analytics Accelerator

After 8-years, working in various roles across the organization, three years ago Keith Berry founded the Moody’s Analytics Accelerator at the directive of the Board. This is a group inside Moody’s focused on launching innovative products typically via a partnership with a FinTech startup.

As many of you are thinking through how to set up your own innovation lab or accelerator, there are a ton of lessons from Moody’s Analytics experience including: (1) how they partner with Fintechs, (2) how trends influencing their thesis and (3) specifically how they’ve approached AI/ML as a big trend influencing so much of their work today.

Show Links


I also hope you’ll consider subscribing to Agile Giants if you haven’t already on:

  • iTunes (also if you feel like the podcast deserves 5-stars, would love a rating on iTunes)
  • Spotify
  • Google Play
  • Or use RSS in your favorite podcasting software


Sean Ammirati: 00:08 Welcome to Agile Giants lessons from corporate innovators. I’m Sean Ammirati, your host, Co-founder and Director of the Carnegie Mellon Corporate Startup Lab and Partner at the Early Stage Venture Capital Fund, Birchmere Ventures. Each week I’m going to talk to guests who are experts at creating startups inside large corporations. I believe fundamentally a startup within a company is the same as one inside the proverbial garage, a group of entrepreneurs trying to make the world a better place using new ideas and inventions. However, I also believe some of the techniques and processes are just inherently different. This podcast is going to explore those similarities and difference.

Sean Ammirati: 00:56 On this week’s episode of Agile Giants, I’m joined by Keith Berry, the Executive Director of Moody’s Analytics Accelerator, a group he founded a few years ago inside Moody’s Analytics. We talk about his background and how the Accelerator works right at the start, so I won’t repeat that content, but

Sean Ammirati: 01:13 I do want to call out three other things to pay particular attention to in this interview. First, I think the way Moody’s Analytics partners with startups for this work is quite interesting and applicable to a lot of you. Second, I think how they think about trends influencing the work going on at the accelerator is relevant. Both industry-specific trends, like commercial real estate tech and then more general ones, like the increasing importance of AI and machine learning. And then specifically on the machine learning AI side, I really love Keith’s concept of an “AI flywheel”. It’s something I’ve been thinking about a lot since we spoke. Many of you are obviously familiar with the flywheel concept from Jim Collins book Good To Great, but applying that principle to AI and machine learning product development I think is quite interesting, and really encapsulates the concept that I’ve observed a lot in some of the startups that I’ve interacted with, both on the corporate and traditional side, in the machine learning space, but haven’t really had a good handle for it or a good simple way to explain it and I think Keith’s concept does a great job of that. I hope you enjoy this week’s episode.

Sean Ammirati: 02:27 Well, welcome to another episode of Agile Giants. I’m really excited to have Keith Berry here with me. Keith is the Executive Director of Moody’s Analytics Accelerator and has a really interesting background that led him to doing that. Keith, that’s probably a good place to start. Could you just kind of walk through your background leading up to the group you’re leading now?

Keith Berry: 02:46 Sure. Thanks for having me on, Sean. So I guess my background, I’d been with Moody’s Analytics for 11 years and Moody’s Analytics is the piece of Moody’s that isn’t the credit rating agency. So we’re a data software and analytics organization. And in those 11 years I joined the organization actually as the head of software engineering and one of my main remits, when I did that, was to implement agile software development as our software development process across the products that we were building. And then moved on from there into a variety of different roles. I’d actually just completed an MBA when I joined Moody’s. And the goal of … I was working in IT within a bank at the time I did the MBA. And a big part of my objective was to learn more about the business.

Keith Berry: 03:42 And interestingly the stuff I really enjoyed on the MBA was the finance. And I’ve always been involved in finance but also the entrepreneurship classes. Then joined Moody’s really as a smaller organization than the bank I was in, because I felt it gave me the opportunity to be closer to that business side whilst leveraging my technical skills. And as time progressed in my career at Moody’s, I moved from building the products to being responsible for a global team of people implementing those products on customer sites. So really a consulting group who went out onto customer sites and made sure that they were happy with the products that I used to build.

Keith Berry: 04:24 And then as the organization continued to grow, I went from there to running one of our product lines. So as the organization grew, we split into individual P&Ls around specific products. And I ended up running one of those from front to back. So that was all the way from product management through the software engineers, but also the implementation consultants. So that kind of full lifecycle of the product.

Sean Ammirati: 04:52 Yeah. So you were a well-known entity within Moody’s when they asked you to start this group. Can you talk a little bit about what the Moody’s Analytics Accelerator is?

Keith Berry: 05:00 Yeah, so about three years ago, I actually think this came all the way from our board. Moody’s is obviously very active in the financial services space and we saw all this disruption coming from Fintech and the large amount of funding that the Fintech organizations were attracting from the VC world. And I think the board was very focused on, “Well what are we doing to stay abreast of this? Keep on top of it.” And my boss, a guy called Mark Almeida, who’s the President of Moody’s Analytics, felt that the right way to approach it was to set up a group who were really focused on a longer time horizon than the rest of the organization. So we’ve been very successful, but like any public company, very focused on the next earnings release and kind of grinding out the numbers quarter to quarter. And he really envisioned a group that was still very pragmatic and still looking to develop real-world product capabilities but on maybe a longer time horizon than the rest of the organization.

Keith Berry: 06:10 So I was asked to found this group and it was kind of a perfect fit for me with my engineering background, the focus of my MBA around entrepreneurship. I spent a lot of time in the organization like I neglected to say some of those roles were in different countries. So I also had good global exposure to the organization. And so it was a really nice fit for me, a dream job, to be honest, to basically go from running a very large organization with a large product with 100 million issue of revenue and about three to 400 employees around the world to being back to square one and creating something from scratch. And so we set about creating what today is known as the Moody’s Analytics Accelerator.

Sean Ammirati: 06:59 Yeah, that’s awesome. And as an entrepreneurship professor, that’s music to my ears. So let’s talk a little bit about this on a couple of levels. So you said one thing is it’s a slightly longer time horizon, kind of what time horizon do you guys think about as your sort of mandate being for Moody’s?

Keith Berry: 07:17 So now we’re in a cycle. I would say that we’re looking to launch new products every year, but we’re seeing that the kind of gestation period of those products is a couple of years. That seems to be the experience that we’ve, what we’ve learned from our experiments so far. So I guess we started the group in September, 2016 and we have about three real products in the marketplace right now. And it’s kind of interesting. One of the areas we dove into has ended up actually leading us to a fairly sizable acquisition last year as we learn more about that space. But I would have said we think of the world as kind of very much in a lean startup methodology type approach, but the early stages are very much what I think of as discovery, and what is it we’re trying to do? And we’re trying to leverage what we can bring to the table.

Keith Berry: 08:24 One of our key goals as we did this was how can we partner actually a lot with the broader ecosystem outside of the company, because the theory being if somebody has invested time and built IP, why try and reinvent it? That seems to make no sense. So we spend a lot of time talking to startups. And I guess when we started the group, I had a couple of key early employees who were also very, had good broad experience across Moody’s Analytics. A gentleman by the name Mikael Nyberg and a lady by the name of Cristina Pieretti, one of them in San Francisco, one in New York, and we were deliberately trying to tap into different locales and locations where we saw startups happening.

Keith Berry: 09:09 It really started with the three of us brainstorming different areas where we felt as a company there was kind of white space and gray space, so areas that we had felt were changing externally on being AI and machine learning. We were reading so much about people doing AI, machine learning, how to we apply that to Moody’s Analytics and products and how could we use that to build products. One being commercial real estate. So we do a lot of work with banks and insurance companies on data and analytics and the amount of commercial real estate assets those institutions hold was growing. We also looked at things like was the more we could do in the asset management space.

Keith Berry: 09:54 But one of the things we really learned from all that was where there’s a lot of venture capital money flowing and private equity money, there’s something that normally that you can dig into and understand, somebody else has got the thesis as well. And we very much found that to be the case in CRE and particularly in the AI machine learning spaces is very crowded right now I would say.

Sean Ammirati: 10:21 Yeah. And I want to come back to that and actually also come back to a couple of things you said there, but first just for people who are probably trying to wrap their head around kind of what this group looks like. So you said you started with yourself and your two colleagues. What’s the size of the group today? And maybe give us a sense on how many ideas you’re sort of exploring at the early phase of that lean process.

Keith Berry: 10:42 Yeah, so we’re at about 35 people today, but we have several products that are now in the market.s And the way we think about it from a funding perspective within Moody’s, and this is a work in progress as well. We very much, my team’s funding is focused around that discovery stage and running what we call POC. So once we’ve got a hypothesis in a space, we’re really looking for who could we partner with, what can we do, how can we bring their … We’re really looking for win-wins, right? Is there a startup with some cool technology that’s relevant to our customers, where we can work together to build a joint product that we can take to our customer base and grow from there? And we will explore that through a series of experiments, proof of concepts. And we will try and typically timebox those to three months or less. And we’ll, as much as we can, limit the funding that they have to about 25 to $50,000. and that we aim to 10 to 15 of those a year.

Keith Berry: 11:57 And our experience has been about a one in five hit rate. So if we’ve managed to do 15 we might have three new product ideas that we would then kind of go to the next stage, which is more of a product development. And we would put some more significant funding behind and start building up a team, sign partnership agreements with the startup or the vendor and really start scaling up. Still experimenting as we go, but very much at that stage we kind of think of that funding a little bit separately.

Keith Berry: 12:37 So I have really, in those 35 people we still very much believe in the core concepts of cross functional teams, kind of two pizza teams. I have a team of about eight people working on a product that we launched called QUIQspread, which leverages AI and machine learning, and they’re working with a partner who’s doing some of that heavy lifting for us. That products just made its fifth sale. We launched it last October and we have a lot of sales cycles in progress right now.

Keith Berry: 13:11 Then similarly we have another product that uses a lot of alternative data in the commercial real estate space called the Commercial Location Score. That’s at that same stage of about 8 people working on it full time. And then I have a group of people that are much more at the early stage at discovery. And then I have a machine learning team, which we can get into why we built our own team, who really are an expert group that we felt as we dug into the machine learning space, we needed some in house expertise.

Keith Berry: 13:44 And then there’s a variety of support functions, from program management cutting across everybody to marketing. And you know, one of the other objectives we have as a organization is, how do we bring that kind of spirit of innovation into the company? So we host a lot of events, a lot of startups that are not only for our group, for the whole organization and really to expose the whole of Moody’s Analytics to what’s happening in the outside world, and more and more we’re placing speakers externally. So trying to get the market to see Moody’s Analytics as a innovative forward thinking company. One of our strengths is the Moody’s brand has been, is over 100 years old and often people kind of, that conjures up the kind of stodgy, slow moving connotation. And one of our goals is definitely to help change that perception externally, and in places internally as well.

Sean Ammirati: 14:47 There’s a lot in there. That was awesome. So I think people now have a pretty good sense of kind of how it works and who your group is, which is great. You talked about VC and PE being one of the leading indicators to sort of develop the thesis of the things that you’re looking at there. Are there any other ways that you’re sort of looking at kind of how you’re prioritizing where you want to focus the efforts of Moody’s Analytics Accelerator?

Keith Berry: 15:10 A lot of customer focus, right? So talking to customers, talking to similar groups. We, as I mentioned before, typically work with banks and financial institutions and quite a few of them have groups that are similar to mine. And so that’s a useful place to understand how are they trying to transform and digitize and where are the areas that they’re struggling. A lot of customer conversations, whether it’s at our events or whether it’s individual customer visits, lead to a lot of ideas.

Sean Ammirati: 15:45 Yup. Nope, that makes a ton of sense. So you’re sort of blending those together and that’s the direction you’re rowing. I will say, having looked at a lot of these sort of groups like yours and I’ve gotten to know one of your program managers pretty well, like you guys do have a much stronger bias to I think partner than a lot of the other groups that I’ve interacted with, which I think is pretty compelling. I’m curious how you think about what to do with via partnerships with a startup versus what to do internally with the resources you have in house and obviously the capabilities that Moody’s can bring to the table.

Keith Berry: 16:26 I think if we can partner, we like that model. It’s a great way, especially in this idea of kind of lean startup and testing ideas out quickly. You can definitely test things very rapidly through our initiative model. We’ve got to be able to create a win-win, right? So we definitely don’t want to distract the startup from where they’re heading. That’s not in anybody’s best interest. And actually I would say one of the things we’ve learned is that the real sweet spot, and this was more a discovery we made along the way, that there seems to be a sweet spot where very, very small startups that are at kind of founder, one or two person stage tend to be too small for us to engage with because there is a risk. You take them in a direction that isn’t in the best interest of what are they trying to do?

Keith Berry: 17:24 And then they’re almost becoming dependent on us and almost that outsourced development shop, which is not the aim. And equally the very large, highly, highly funded, hundreds of millions of dollars, hundreds and hundreds of people start up, their difficult actually for us to partner with as well because we’re just another sales opportunity for them, to be completely frank. And there they have a lot of momentum and there’s some sort of sweet spot and I think it’s somewhere in the kind of three to five million dollar revenue range typically that we see, where maybe there’s 30 to 50 employees but we’re still engaging with the CEO or the senior level and how is it that we can work together and create that sort of win/win. So that’s definitely been the sweet spot we found.

Sean Ammirati: 18:18 And how do you manage expectations in those kind of … So I think a lot of executives in roles like your with think, “Yeah I want to partner with startups that look like that,” but they end up with both sides kind of frustrated. Whereas I think you guys do a good job finding that win/win and making sure both sides understand what’s a win from the other person’s perspective.

Keith Berry: 18:39 Yeah. I think a lot of it is actually relationship and building those relationships and building that trust. I would say, I think a lot of the startups are often quite cautious and probably rightly so, of the big corporation. And there’s definitely steps in the process of working with a big corporation where it can feel, we’ve got to manage those steps well from our side and from their side. And thinking about some of the engagements with legal on a legal contract writer, the legal firm with the legal side of a large organization is often set up to deal with large contracts with big outsourcing our system, integrated providers, the Accentures or the TCSs of the world and not so much a $25,000 contract with a startup of 20 people. And kind of holding people’s hands through those process of making sure that the startup, does it get offended by the legal tone. Sometimes the legal organization understands this is a limited experiment that might grow into something else. A lot of that is is just managing the relationships and the personalities as we work through.

Sean Ammirati: 20:07 100%. So you mentioned that you have your own internal AI machine learning team, and have already mentioned that that’s a big trend you guys are focused on. That is one of the areas where there’s a ton of opportunity, but often also confusion about what’s possible and what’s not possible. How does that influence the stuff that you’re building around this mega trend?

Keith Berry: 20:32 Yeah, so really a couple of things to unpack there. We started with some discreet use cases that we thought we could apply AI and machine learning to, and then started to basically reach out to startups. So one of the first ideas in the world that we live in, there’s a process called financial spreading. And this is where you, if you’re a customer of a bank and you go in and ask for a loan as a small business, they’ll ask you for the last three years of your accounts and you’ll typically give them a pdf or a printout from your accountant. And many, many banks will do a process that they call spreading, where they manually key in that data. Or almost everybody manually keys in that data. And we actually have some internal use for that. But also it’s a big thing we know a lot of our customers struggle with. Actually leads to some of the long lead times that you see on banks approving small business loans, is just some of that manual process.

Keith Berry: 21:34 So we felt that was an interesting challenge, because we had some data assets in machine learning, you need some training data and you need some data to test with. And we from our own internally solving of the problem had had some good tests. So we reached out to a lot of startups. And what I found in that space is everybody says they can solve the problem. And in fact as we dug and dug further into that space, we found that very few of the firms we were talking to really could. And there’s quite a lot of smoke and mirrors I would say in the AI machine learning space. And that’s partly what led us to our first hire, a gentleman called Ashit Talukder who we were lucky to hire, who was very experienced in AI machine learning. And we really initially hired him to help us with due diligence of the various startups we were talking to and really understanding what were they really doing with AI and machine learning, and how are they leveraging it.

Keith Berry: 22:43 And I think what we’ve discovered is, I am somewhat skeptical of the very general purpose AI machine learning tools that are out there, that all try and … There’s a whole suite of tools that set the vision that a business analyst will be able to feed in some data in and out will come this amazing algorithm that will solve all your problems no matter what that problem is. But I think in the AI machine learning space, if you can define a narrow enough problem space and you have some unique data, or you can generate that data through a workflow, you’re really well positioned to build a really interesting competitive mode, because the training data itself becomes the competitive advantage.

Keith Berry: 23:36 And the way I think of it, I think Tesla actually explained it really well. And I’m not sure if anybody listing has watch their Autonomy Day, which was a fascinating thing where they tried to explain that thoughts about self driving cars to stop mark analysts was a fascinating challenge. But they were talking about how, as they put more and more Teslas on the road with the self-driving capabilities, they’re uploading significant amounts of data on the driver inputs and really they’re building a better and better neural network to solve that self-driving problem. But they’re learning from what the humans behind the wheel are doing. And so there’s kind of an exponential growth as you get more people driving the Teslas, they’re getting more feedback and the thing becomes self reinforcing. So I think there’s some really interesting product design things to think about to create that flywheel. We think of it as a machine learning flywheel where you encourage the users to basically help train the machine, and the machine gets smarter and and the user input becomes more and more just small corrections. I think it’s having a narrow enough space that makes that work well.

Sean Ammirati: 25:04 Right. And having the humans in the loop. I like that term, machine learning flywheel. I often tell people, I think the human computer interaction department at CMU often has more, like some of the faculty that I interact with there have more interesting insights for me as I think about the future of this stuff than even people doing hardcore AI and machine learning today, because making the interface clear enough that the users understand and are able to give you feedback, or in the case of like a Tesla where it’s just sort of built into how the product works and there’s training data coming back, really does do that.

Sean Ammirati: 25:36 I’ve never thought about it as sort of taking that Jim Collins flywheel concept and applying it to machine learning. But I like that a lot. That’s good. So if you were to contrast that then with, for example, your commercial real estate opportunities, it seems like you have to take a slightly different approach for innovation there versus in some of the other spaces because of sort of how it works and also maybe some of the high first reality opportunities there. Like is it safe to say you guys probably wouldn’t build the equivalent kind of group for some of these more horizontal stuffs like you’ve done for machine learning? Or do you think you’ll end up building out like a commercial real estate group of expertise and other kinds of hot button vertical or horizontal areas?

Keith Berry: 26:20 In the commercial real estate space, we have made a fairly sizeable acquisition, which I guess has brought some of that skill set in house. So, I mean that was really opportunistic and I think it comes from, as you explore the space and you learn more about it, lot of M&A opportunities come your way as well and it’s a bit of the same. Why would you try and build everything if it already exists out there and somebody is actually looking for an exit?

Sean Ammirati: 26:50 Is it public, the acquisition that was made?

Keith Berry: 26:52 Yes, it was an acquisition. It was actually a small public company called Reis. We closed on that last October.

Sean Ammirati: 27:00 Did you come across them via the partnerships type model or was this a different sort of front door into the Moody’s Analytics Accelerator team?

Keith Berry: 27:10 It was really as we explored the commercial real estate space, we noticed there was a lot of VC and private equity money flowing into the space and very focused around prop tech. And really, I think everybody’s philosophy is very similar, which is commercial real estate asset class is one of the largest in the world, but very opaque with, there’s not a lot of data available in that space or easily available. And the data is very much, it’s very slow moving. So understanding your current position, your current exposures is challenging.

Keith Berry: 27:49 And as we explored the space, one of the things we often do is map out the market. So who’s in the market? And Reis was definitely one of the companies in the market, kind of small to midsize player in the market, but with a good data set and a good reputation and a good customer base. They decided to put themselves up for sale, I guess, kind of early Q2 2018 and we took part in that process and ended up being the winning bidders.

Sean Ammirati: 28:23 So I want to be respectful of your time, but just a couple of last questions here. So maybe if you could give an example of a successful project that you worked on and some lessons from that, and then maybe something that didn’t work out in these first couple of years and what you’ve taken from that as well.

Keith Berry: 28:40 Yeah, I would say in terms of successful, the QUIQspread product that I was talking about, now it’s launched and truly getting some traction. We really got that into the market in about, from idea to launch, in about 18 months. It’s just slightly longer sales cycles. So we’re only just starting to see those sales come through right now. And I think it was just a, it’s a good example of many of the things we’re trying to do to be honest. We built it with honor, a lot of customer involvement, a lot of iterations with customers. As we’ve gotten, started having customers actually providing training data. Had a very joint development sort of stage as well. And that’s evolved quite nicely. One of the things I would say we’ve learned from that is within … We’ve kind of added our own business development people to the organization as a kind of key role on the team.

Keith Berry: 29:44 So when I think about the makeups of some of our product teams, with these very early stages… So one of our greatest strengths at Moody’s Analytics is our customer base. So pretty much every financial services organization buys a product from us, so we’ll have a relationship with them. And that’s obviously one of the things we try and leverage as we launch these products. But you can’t expect our sales force of, I think we’ve got a global sales force of about 750 people around the world, to understand and be able to talk to customers about sort of the new stuff we’re doing. Just because it’s too early and it’s changing so rapidly.

Keith Berry: 30:22 And so we’ve really added our own layer of business development within each product. So a typical product team will have the founder or the head of the product and then the person, kind of right hand person reporting to that founder is typically the head of business development. And now you’re literally demoing three, four times a day and getting feedback. But keeping that feedback really close to the product team, so that we can make those really rapid early stage iterations. And so that’s actually worked out really well.

Sean Ammirati: 31:01 That’s really interesting. How much experience did those BD folks have before they take this role within one of your projects? Like are these people right out of school or are these people with some work experience? What does that work experience, if they have?

Keith Berry: 31:17 They’re definitely people with some work experience. I mean ideally, some domain knowledge, right? Because they’re going to be talking to customers about what the customer does all day. So in the real estate team, for example, we have a couple of people in that business development role and they both worked in the real estate side of the business themselves. So they can go and have a real world conversation with the customer and kind of display credibility, but at the same time understand the feedback very quickly. So I think that’s the ideal, but still a very open mind towards product development. And one of the keys we found we were keeping the small teams is people have got to wear different hats compared to a more traditional team within our organization. People are wearing multiple hats and trying to be like a startup and rolling up their sleeves and doing whatever they need to do to kind of make the next release of the product work and get the product out there.

Sean Ammirati: 32:25 Yeah. So one last question on this and then I’ll ask you a final parting question. But you mentioned the sales cycle’s a little longer for this product. How did that end up impacting the POC stage between you and the startup you were partnering with? Because you would think it’d be harder to get that proof with the long sales cycle.

Keith Berry: 32:44 Yeah, actually that side of it hasn’t been so bad. I think it’s interesting. I would’ve said the bigger challenge with the longer sales cycle is keeping the momentum within the organization. So you’ve launched a new capability and people are excited, that there’s almost, I find, if something doesn’t sell immediately people are, “Well why hasn’t it sold immediately?” So it’s kind of keeping that positivity around a new product if it is just a longer sales cycle. And in the example a QUIQspread, organizations are, it’s a kind of strategic change. They might be doing all of this work manually today and we’re certainly not replacing all the people, but they are looking to do it because they want to make efficiency savings and speed savings in the processing that they’re doing. And so having to think through all those ramifications.

Sean Ammirati: 33:44 Sure. I guess when you did your POC with them then, were you doing that with internal Moody’s processes then? Or did you, were you able to get customers to sort of-

Keith Berry: 33:56 We were able to get customers to provide us with test data and we would do weekly scrum calls to demo to a couple of early customers where we were, what was going well, what we were struggling with. So a very transparent process with the customers.

Sean Ammirati: 34:16 Cool. Well, I want to transition for this last question. Think back to when you were just finishing up your MBA and joining Moody’s and had this entrepreneurial interest. As you know, that’s a lot of the students who I interact with, they’re getting their MBAs and they want to be entrepreneurs, but often are taking more jobs with public companies to do that is maybe smaller than where they’ve come from, but still established companies with a long history.

Sean Ammirati: 34:43 I’m curious what advice you would have for a student like yourself coming out today who wants to have a similar impact that you’ve had and opportunity to ultimately grow to run a group, found and run a group like you’re doing inside of Moody’s?

Keith Berry: 34:56 Yeah, that’s it. Well that’s a great question. I mean I think the good news is organizations are generally, it feels, moving towards a place where they’re looking to act more like startups. I think it’s just, we’re part of this corporate entrepreneurship community group who is a whole range of corporate startups including people like ourselves, Vanguard, Proctor and Gamble and Microsoft, are trying to apply the ideas of the lean startup to their businesses. And I think people have realized that the great way to innovate, and at the same time reduce risk, right? So kind of moving away. I think there’s generally a trend to move away from their five year projections that lead to you investing $10 million up front in a new product that never really gets anywhere. So this kind of rapid experimentation is definitely, is applicable to any product team and anybody in the product and even solving internal challenges within a company.

Keith Berry: 36:14 I think that’s one of our biggest takeaways from running the group, is starting with small experiments we found can change the whole organization. You know, interestingly our group was the first group at Moody’s to start using Slack as a collaboration tool. Partly because we were using it with a lot of startups. We’re now one of Slack’s biggest corporations, because it kind of grew out from our team. I think taking … Not abandoning the … If you’re starting in an organization, how can you kind of stay up to date on those ideas, things like the lean startup, there’s a lot of entrepreneurship books and podcasts and things you can follow now, and how can you apply them to the situation you’re in. And I think organizations are becoming more and more open to that and from there I think there’s definitely some luck.

Keith Berry: 37:09 I think it’s interesting the way things seem to, if that’s your interest, how you can keep coming back to it. And now that would be my advice is really stick with it. Keep on top of the latest thinking and some great conferences, great books, great blogs and podcasts out there and find that champion in the organization who’s got the interest that you can educate and try and apply some of this stuff. And one experiment at a time, I think. Find a little project that you can try it on and then go from there.

Sean Ammirati: 37:47 I think that’s one of the real transferable things between, the lean startup methodology for two people in a startup accelerator and the lean startup methodology inside a place like Moody’s. I know Eric [Ries] is part of the CEC that you guys are part of as well, and I think that that’s well said. And it’s just been really great to learn more about what you guys are doing at Moody’s. Thanks for the time today, Keith. I really appreciate it.

Keith Berry: 38:13 Thank you. Great conversation. Thanks.

Sean Ammirati: 38:15 Awesome. I hope you enjoyed this episode of Agile Giants. If so, consider sharing it with a friend, and if you think it’s worth five stars, which I hope you do, please go to iTunes and rate it so that others can find this content as well.

Join the discussion

More from this show


Episode 19