How Bain’s Micro-Battles Drive Agile Digital Transformation

Tom Soderstrom:
Welcome to the Executive Insights podcast, brought to you by AWS. My name is Tom Soderstrom and I'm an enterprise strategist in AWS. Today we have the pleasure of having Frank Ford from Bain Consulting with us. Welcome Frank.

Frank Ford
Thank you Tom.

Tom Soderstrom:
Would you talk a little bit about what Bain does and what you do for Bain?

Frank Ford:
Yeah, absolutely. Bain and Company is one of the leading management consulting firms and we help clients with a very wide range of challenges. At Bain, I focus on technology topics. I lead our cyber security practice, I lead our cloud practice, and also do quite a lot of work around technology strategy and basically helping companies drive some of their key priorities using technology.

Tom Soderstrom:
Interesting. One of the insights we have… It’s not really an insight, it's an of course, technology is useless unless people adopt it and bring it forward. How do you deal with adoption strategies? Let's start with the executive level. How do you go about it? And then I'm going to talk about what I see the trends are and see what you think.

Frank Ford:
If people ask what is management consulting really about, the straightforward answer I think is helping clients to make change, because it's not about the analysis or the strategies, they’re necessary, but it's really about helping clients drive through change. One of the principles that Bain was founded upon back in the day was it's about results, not reports good. It's one of our taglines that we've had ever since the company was founded 50 plus years ago. Now how you do that really depends on what the topic is, but some of the key elements are making sure there is very clear sponsorship at the top of the organization. We refer to something that we call a sponsorship spine, which is what are the key people that you need aligned around the change through the organization to help you achieve the real change? Because top-down sponsorship is often essential, but it's not sufficient. You'll need the various layers of management all the way down to the people who are actually doing the work to understand the reason for the change, why the company's doing it, what their part of it actually is, to make change effective. That's one principle that we find is very important amongst others.

Tom Soderstrom:
The executives that we've talked to, I'll give you what their priorities are, and that’s changed this year. One of the things that's emerged faster is the need for speed. The need for speed to profitability, to change, to compliance, to bring in skills. It's really all about speed. What do you hear that are priorities from the customers you talk to?

Frank Ford:
Very much would agree that speed is a critical topic. I think the reality is that the evolution of technology in recent years and the continuing evolution is speeding things up. The ability to make quite significant impacts or changes in the way that you engage with customers is different now than it was even five years ago. The cycle time is just reducing across all aspects of business and the public sector as well. The reason we're all seeing this is really driven, I think by technology, and it's just a requirement for companies to be able to go faster than they currently are.

There's some interesting observations around this in terms of how companies change. We're back a little bit in time, a few years maybe, companies really thought about large-scale transformations if they were looking to make radical change in their company. Even today, about a third of companies are undergoing some type of large-scale transformation. When I say transformation, I mean something quite radical. Not small-scale things, but a major change in the business. About a third of companies are doing that today, and if you look at over the last three years, most companies will have tried two major transformations on average. The failure rate though is really, really high.

Tom Soderstrom:
Do you have a number?

Frank Ford:
Yeah. The companies that claim they're fully successful with their transformations is about 10 to 12%, which is a really low percentage. The companies that admit to failing completely is also about 10 to 12%. And then you've got the middle band of about 75%, where they partially succeed. They accept diluted results, but they didn't really achieve their ambition. The old method of driving change and big structured programs for major things is not very reliable. The odds are pretty low.

So companies are starting to move a bit in terms of how they think about change. Moving away from the large structured programs with very slim chances of success, and moving more into methods of continuous change rather than separated change. Also, thinking differently about how they communicate the goals of the change. Less about target-driven, more of an aspiration-driven. You share the vision and the aspiration, help people understand it and buy into it and why it's so important for the company, and you'll find people will start to work it out themselves as well as part of these more continuous change approaches.

Tom Soderstrom:
It mirrors what we're hearing. If you really want to transform a company, the most effective ones that we've seen is you don't create one excellent ship, you create the tide that raises all ships across the business. But if that's true, how do you do that? How do you get the entire enterprise to buy in? Do you go at that level?

Frank Ford:
One of the things that we talk to clients a lot about, and this is something we had a brief conversation about before, it's this concept that we call “micro-battles”.

It's a way of making impactful change in companies. And so the basic idea is firstly you need to choose the right topic that you want to go after. The idea is you don't start small with things that don't particularly matter, you actually start really big. Your two or three toughest problems. You work out what they actually are, you identify. It could be, for example, and I'll talk about an example with a consumer products company that we worked with on these topics. Their toughest problems were really about driving sales for some of their key products. Very straightforward, but a major focus for their business to drive growth for key brands. You choose the right topic.

And then you want something which is winnable. Now what we mean by winnable is, it's something you can actually make real progress on in a relatively small amount of time. The window is, at most, three to four months. And so what we mean by that is you can develop a prototype or a new way of working, actually get a solution. You can crack the problem within that type of timeframe. Now, it can be shorter. This particular company, they set a four-week window for their particular topics. You don't want it to be much longer because it's hard to sustain the energy and the focus and the enthusiasm if you expand it too much. The final part of it is, it's something that you can scale.

The framing of micro battle is that micro being, you can do it in initially a small scale, but then you can actually scale it out globally across the organization. A very tough problem, something you can crack within a no more than a three to four month period, and then something that once you cracked it and you've proven it works, you can actually scale it across the organization. It's a way of driving change, which doesn't require you to mobilize lots of people, lots of resources, lots of buy-in. It's got to pick a topic, solve it in a country, work at how you can extend that and scale it beyond that country, for example.

Tom Soderstrom:
Let's talk about cybersecurity inside the micro battles. I was NASA's first chief technology innovation officer at Jet Propulsion Laboratory and it took me a while to figure out that I got to bring in cybersecurity early because in the end, cybersecurity can protect us from doing dumb things, but also, you need cybersecurity's buy-in to go live later. How do you deal with that in these micro battles?

Frank Ford:
I'd agree with the basic idea that cybersecurity needs to be involved very early if you're developing something that has any implications for cybersecurity, anything to do with technology, for example. There is a real cost of bringing in cybersecurity late because what often happens is either something's been envisaged or designed, developed in a way which is not adequately secure.

So then either you have to go incur more costs and make changes. There's generally great reluctance to do that because it slows things down and often the resources aren't available.

There are many organizations that feel cybersecurity is a barrier. It slows them down, it creates problems. We've seen some organizations where when someone wants to do something they have to fill out lots of documents about the solution, the approaches, how they're going to solve certain problems. One company I saw, they had 27 different forms to be filled out before you could start anything.

The key evolution for cybersecurity teams is how to actually help the development teams and to make it as easy as possible. Some of the key ways that companies are starting to do that is training their people and deploying their people to work as part of these two-pizza teams.

Tom Soderstrom:
Exactly.

Frank Ford:
It's helping people understand what needs to be done from a security point of view, working with them to find the right solution rather than being dogmatic that it has to be done a particular way. It has a very positive effect because developers, in our experience, are not opposed to building good cybersecurity.

Tom Soderstrom:
But they're opposed to filling out 27 forms.

Frank Ford:
The forms are not very popular, that's for sure. When they understand it better, then they are quite willing, and often if you talk to them early, they will create with you and come up with solutions.

Tom Soderstrom:
Completely agree. So, one of the things we saw, I'm very curious to see if you see other companies do this, was instead of, because security was always a delay, if you can't go live, if you're not compliant, it's costing money. Instead of the cyber security people run a script to figure if it's compliant, give that script out to developers, they run the script and now they get a more secure solution and it shortens that time. That helped us at NASA. How does it work for your companies practically? What do they do to speed up this waiting for compliance check?

Frank Ford:
Some companies are very good at this and so the ones that are very good at it have really focused on driving as much automation into their software development life cycle.

Tom Soderstrom:
Yes, that's right.

Frank Ford:
Which is not just around writing the code, but it's also about the cyber security checks as you go along, so it's built into the development process. For example, relatively early on there would be automated tests against the code to see whether it meets certain requirements, has certain vulnerabilities, uses latest patches, et cetera, and compliance checking and so forth. It's built in. It's very difficult actually to produce code at the end that's not compliant because of the way it's all automated and built through. Setting one of those things up is quite an endeavor.

But the target state is this more automated, more inbuilt machine to develop code which automatically helps people develop it in a secure way.

Tom Soderstrom:
I wish that I'd had, in those days when I was a developer and software leader, the tools we now have. I'm really happy about Amazon Q Developer, and the other cloud providers are also having coding assistants. It's really helpful because you can bring in compliance and security upfront. I think we will see the level of innovation go sky-high because of these automated processes that delay us and people don't want to do and they insert vulnerabilities in our code. If they could be automated, what are we going to see?

Frank Ford:
Yeah, I agree. I think the basic point of the cycle time for developing prototypes and code into production keep reducing, and I think we'll continue to see that and I think gen AI is going to help a great deal as well in terms of generating code samples. There's more and more automation coming in.

Tom Soderstrom:
When you talk to the executives and presumably the teams, is there a big worry that AI will take my job? Are you hearing that or are you hearing the executives saying, "Yeah, I'm going to cut my staff in half because I have AI."

Frank Ford:
I think it is a concern in some industries and I think to my mind it's just another example of automation and the impact of automation. Basic robotics and manufacturing on assembly lines has been in place for decades. That changed the nature of the jobs in those manufacturing plants. There are still jobs but they're different jobs and there may be less number of them, but the skill requirements changed radically. Generative AI will have a similar effect for some jobs.

I think most companies are looking at it an assistant, as opposed, at this stage, versus an outright replacement so that there are some tasks that can be automated using AI to help people focus on the higher value activities. If it's, say, dealing with a customer's inquiry, the ability to deal with it to a higher quality more rapidly through AI and actually really help the customer so they feel happy, is a good thing and the people can intercept the calls which are more complicated or require a person to talk to.

Tom Soderstrom:
One of the key worries is I don't have the people I need. We tend to disagree. We think that you do have the people you need, they may not have the skills they need. Now what you can do is you can solve it by just up-skilling, because if you just replace all the people, you lost all the business knowledge too.

Frank Ford:
I think it's a topic that comes up very often across many different domains, and so companies do need to get quite thoughtful and serious about the capabilities that they need in their business and how they acquire them and have the right strategy for acquiring them. Some of it will undoubtedly be retraining and up-skilling people. Also at the board level as well. There's lots of issues at the board level in terms of the relevant skills, particularly around technology topics. Also outsourcing other sourcing methods to actually build the skills that you need into the organization. It's not necessarily that you'll have to have everything within your organization. There are certainly skills that you can and should outsource, and most companies look at that as more commodity skills. They will outsource the ones that are really important for their core business. For example, in technology terms, it could be around cyber security, it could be certain engineering skills, it could be architecture for certain. Those types of topics we want to keep in house.

Tom Soderstrom:
I really agree with, and that's one of the key things, it's so tempting, the board of directors says you need to do Generative AI. Now, okay, I'll outsource that. Well if that's the future, don't outsource it, bring in experts to help you. But you've got to grow, and then get rid of some of those things that are not differentiating.

Frank Ford:
I think it's very important, I think when you recognize something like AI is going to be an increasingly vital part of everyone's business as time goes on. What drives AI is your data, your data architecture, your data strategy, so you really want to have your skills in-house for how you really understand your data and how you architect it and make it usable for your core business processes and for generative AI. Those types of skills are extremely important for companies to retain. They're very differentiating.

Tom Soderstrom:
They are. About data, like you said, data is key and it's one of the key trends we're seeing. What strategies do you see around data?

Frank Ford:
It's a great topic. I think too much lack of strategy, I think, is what we see around data. What I mean by that is companies need to take it a lot more seriously and really think hard about how they're going to structure and use the data. Who owns the data? Who are the stewards, basically. How do you ensure the quality of the data? Where do you actually put the data, how do you architect it and structure it so it's usable and accessible? Simple topics to describe, but actually quite challenging for companies to solve in practice. But there are lots of best practices that companies can deploy. For example, in terms of where you have data stewards in your company responsible for the data, where companies generally end up when they've gone through a journey of maturing, how they think about data, how they use it, is maybe a small central team in terms of standard strategy policy around data. But the ownership of the data sits out in the business, with the business and specialists who understand the business and the data in designated roles, and you have a hub and spoke type arrangement.

But companies often go through a bit of an evolution because they tend to have a lack of people that really understand data, lack of data scientists. They often start off with, they're dispersed all over the place. They pull them into the center to try to get control, but then that removes them from the business, which is where you want to use the data, so they eventually, normally, end up in a model where you have a small hub and then in the business managing the data and helping the business use it effectively, which is the key thing of course.

We do see more companies starting to address this topic. I actually think that gen AI is going to really drive a lot of hard thinking about data. What we're seeing is companies pursue gen AI use cases, but use case by use case. Not trying to work out what they should do with the data to help them enable use cases at a lower cost. We think people will start trying to drive value through gen AI, piloting things, et cetera, one-off, not restructuring their data, just solving the problems as they go tactically. But then we see them increasingly reach a stage where they're going to take a step back and realize to go faster, they're going to have to take some time to sort out their data first.

Tom Soderstrom:
One of the things we've seen, the language you use really is important, and data owner is a challenge because the data owner can... We had to do this with many company, is actually you are not the data owner, the company is the data owner. You used the term data steward, which I like a lot, but not that many companies are using that term yet.

Frank Ford:
It's interesting. I think there are two key roles, just to pick up the point briefly before I answer your question. Data owner and data stewards, they're separate roles. The way we think of it is that a data owner is the usually quite senior executive that is responsible for the data from a corporate perspective, basically making sure that the data is collected and usable by their business to drive their business goals. The data stewards are the individuals in typically full-time roles whose job it is to actually make that happen in practice. Make sure the right data feeds feed into the right place, the data quality is okay.

They help the business work out how to build whatever dashboards they need, for example, to drive minimum management decisions related to the data. Their job is not just to collate and collect data, it's also to help the business actually use it to drive business value. Typically, as part of the management team. The data steward is a very important role for businesses. And then your point around sharing data and making sure it's accessible. I think the best way of doing that is making sure there is a senior leader in the business that has the influence to drive the right behaviors. Companies increasingly adopting a chief data officer that would report to a very senior level in the business. The typical reporting line, it may be the CEO, it's often the COO for a company.

Tom Soderstrom:
You're saying usually the chief data officer reports to the COO?

Frank Ford:
Yes. Where they often report today is one of the two, and I think it's more often the CIO today because data is seen as a technology topic. But increasingly we're seeing businesses think through what it's actually more of a business topic, and so to make sure that the business is really thinking about how it utilizes data and drives value from data. Having it reporting into a business leader, the COO for example, takes it away from being a purely technical topic and you're enabling the people to use the data by building platforms. But that's not enough, you've actually got to drive business adoption, you've got to drive the right thinking about business use cases and so on.

Tom Soderstrom:
The data scientist was a role I created at Jet Propulsion Laboratory and CTO community. In the beginning, everybody's scratching their heads, they get the title, now what do we do? One of the new roles is Chief AI officer and one of the new variants of that is Chief Data and AI officer. I'm seeing that pop up. What do you see? Because you cannot create a chief-next-new-thing-officer all the time. Where is the landing in AI as far as what you're seeing from executives?

Frank Ford:
I think the jury is out a little bit still. I think the businesses are very much asking themselves the question, how big is it really going to be, the AI and gen AI in particular, to really help us drive change in our business? How fast is it going to come? Because in the early days, early days being a year ago, companies were being told and being convinced that there'd be a 40% cost improvement in some of the core operations, which is massive, right? No one's really achieved that.

But that was the hype cycle, so then a bit more realism has settled in and people start to understand what it takes to drive real scaled value through these technologies. The jury's out in that way, I think people are now getting more realistic about how to use the technologies and drive value. There still aren't many very scaled use cases fully enabled by gen AI. There's some starting to emerge and we're working with some clients on these topics, and there's some very interesting ones. Marketing is a great one in terms of automated content generation at very fast pace. When you get the quality right, you can get messages out about your brands in near real time and you can link it to what's happening on social media or whatever else, what people are interested in, and you can hit people with the right type of messaging in real time.

People see it as being real and something to invest behind. We're seeing more of those emerge now. But your question about how companies are organizing around it, they're still working it out in many cases.

Tom Soderstrom:
You're not seeing a trend yet?

Frank Ford:
Not seeing a clear trend. I think that they will have normally a technology leader focused on gen AI and helping the business, enable it and drive pilots and this type of stuff, and companies will set up programs and so forth. They'll collaborate with tech firms to help them accelerate the pace of the work and consulting firms as well. But in terms of what type of major org changes is it driving, I think it's a bit early for that yet.

Tom Soderstrom:
What are three things you would recommend to a new executive who was just elevated to any role in the executive staff?

Frank Ford:
I think understanding technology and what it really means for your business, what the opportunities are, is very, very important for senior executives. You touched on it before, but we do see an issue where boards are not sufficiently understanding of technology and it's actually very bad for the company. Companies that have properly educated boards that really understand these topics are much more effective at adopting change and driving technology-based improvements in their organization. That is one. I think finding a method to accelerate meaningful change in the company is another, where we started the conversation. Whether it's micro-battles or some other technique, but you need a method that is actually going to be successful for your business. You go back to the basic stat that only 12% of big transformations work and the other ones either fail or have diminished value. That means a lot. It means that-

Tom Soderstrom:
Yes, I agree.

Frank Ford:
... seven out of eight companies that try these things are not going to succeed very well. A lot of time the money wasted, opportunity, and it worsens your market position versus those that actually succeed, so cracking that problem is really, really important for companies. And then just picking up on one of the things that we talked about in terms of capabilities.

It's amazing the number of times that issues in businesses, performance of many dimensions, come down to lack of capability of one type or another. Most companies do not have a well-thought-out capability map, so what do they actually need to succeed in the business today, what do they need tomorrow, and how are they building that in a meaningful, cost-effective way? You ask a company if they've got one and ask them to show it to you, then 99% of the time you won't have anything and people will be trying to solve it in different pockets of the business. But there's no big picture, there's no sufficient concerted executive effort too often on those types of topics. Those are a couple of things that I think are very important for companies to focus on.

Tom Soderstrom:
No, thank you. I like that. I would agree with all of them. We think that, or at least I think, the culture of experimentation where you can actually experiment, your micro-battles, it's the same idea. Try things and the things that are business value, then evolve them. The other things, just park them. Fast to speed. So thank you Frank. Thank you so much for joining us.

Frank Ford:
It's a pleasure, Tom. Thank you very much.

Listen now

Listen now

Listen now