Virtual Summit Recap: The Future of M&A

Hello M and A friends. This podcast is brought to you by dealview. M&A is coming back. Would you like to avoid wasting hundreds of hours managing Excel Trackers and building PowerPoint reports? Do you want to make your process more efficient? Well, I have great news for you. Dealroom is MNA management software to save you time, headaches and improves collaboration so you realize value faster. Learn more@dealroom.net here's to the deal that was my lovely daughter Shyla. If you want to keep up with the latest on M and A Science, our live streams networking events. When we release new resources, visit mascience.com and subscribe to our newsletter. You can Also follow our LinkedIn page to catch highlights of our interviews and if you like our content, please tell a friend so we can keep growing our community of forward thinking M and a practitioners. The HR M&A Roundtable's annual conference has gone virtual for 2020. This world class event focuses on one of the hardest parts of M and A the human side. Join me in a great lineup of expert practitioners as we share how to deliver transaction value through the people, leadership and culture elements of a deal. You can save 10% on your registration when you visit M&Roundtable.com and use the discount code Science. Now on to the interview. I'm Kisan Patel and you're listening to M and A Science where we talk with deal professionals and learn valuable lessons from their experience. This podcast focuses on stories, strategies and what actually happened during M and A deals. This interview is a rewind from our virtual summit. Hope you enjoy it. If you want to see the full version with video, visit mascience.com welcome everybody. Today I'm here with Dr. Karl Popp, Senior Director of Corporate Development at SAP. Karl is a board member of the German association for Post Merger Integration, member of the German association for Mergers and Acquisitions, and chairs the Program Committee of the European Workshop on Merger Integration. Today our talk is about the future of M and A. We're going to be discussing what tasks are automatable in the M and A process. You feeling good about this?

Yeah, absolutely. It's one of my most loved topics and again, thank you for having me. I think it's a great way to share experiences and best practices. I love it.

If you want to kick off, Dr. Karloff, could you briefly describe your role at SAP?

Yeah, so I'm part of the Corporate Development team at SAP, headquartered here close to Heidelberg in Germany and I'm running product due diligence and post merchant integration projects and I'M also working on the definition of the M and A processes that we use at SAP.

So, first question, what is the future of M and A?

Yeah, I think the future of M and A is something that's a lot different of what we see today. Automation will play a major part in massive changes that will see a lot of the, I would say, annoying craftsmanship that you see today will be done by software, automats and robots. So we'll have much more time to focus on mitigating risk solving, problems planning, integration and such.

How do you come up with the idea of automating M and A?

Yeah, so I work at a software company and I see a lot of great innovation around me with predictive analytics, natural language processing, machine learning, robotic process automation. All the buzzwords I see live happening around me and I go back to my workplace and there's nothing of it. And that kept annoying me. So I started to work on automation of MA processes.

What areas of M and A do you specifically look at for automating?

We basically look at the end to end M and A process, or from very early strategy to closing the integration project. I've also created a new book which will be out in German at the beginning of July and in English at the end of July, talking about what are the tasks in the M and A process, how can they be automated, which data are used, and that could be a basis for further automation.

When you look at that ability to automate some of these tasks, are there areas that you think are generally easier than others? The sort of starting points?

Yeah, absolutely they are. The best starting point is a data lake where you put some automation on top and basically you could think of a data lake like the data room. And there are companies out there who already offer certain automation on top of data rooms, also some analysis of documents, analytics, et cetera. So that's the typical starting point where you have lots of data and can easily start to apply the household machine learning, analytics and other tools.

What's the timeframe? Do you see this? Because we're starting to see there's data rooms and you're starting to see the legal industry using contract analysis tools. But there seems to be a pretty big gap from that era to the regular practitioner that's coming in and typically doing diligence. What kind of timeline are you seeing before we see a big shift in this area?

I think it will take another two to three years, not longer, because competition will drive the technology forward. I would say the competition for just doing better and more accurate Due diligence based on this automation that you can use. Examples in the software business are automatic tools to analyze licenses of open source components. And there are even tools who can do things that only architects could do, which is based on source code, analyze the architecture of software products. So there's a lot of innovation happening. I think the key hurdle we have today is bringing all of that together. Because in the end, to end automation, you don't want to have islands of automation everywhere. You want to have a joint basis to integrate all the automation that is available. And that will take some time. The race is on.

I would say just rough estimate. We're looking at a 5 year, 10 year, 20 year timeline.

I've been a scientist, so I would say it's more like a guesstimate. Say three to five years from now will be a different game with a lot more automation.

When you talk about creating a data lake out of the data room and starting to do some of these more of the automated analysis, can this actual data change the M and A approach itself?

Absolutely, yes. So I believe there will be a shift of the human capacity in M and A processes away from searching and analyzing into mitigating, taking decisions, etc. If you look at other parts of automation, say you have four weeks of due diligence and you usually wait two, two and a half, maybe three weeks for the results of the legal analysis from the legal advisors. You could have a draft version based on automated analysis today, already on day two, with the major risks automatically found. And then it's a whole different game. You have almost four weeks instead of one and a half weeks. Assume that you have all the documents at the beginning of the due diligence. So it's a completely different thing that we are walking into.

I'm curious if we can define automating a task because are we looking at some of the contract analysis tools essentially give humans the superpowers to really be able to be more productive. Maybe if you can just share your approach to automating processes in M and A.

The area of analysis and even construction of legal documents is an area where we are very far advanced in applying modern technologies. So there are tools out there who can really do a very fast automated analysis and identify certain risks, I would say, but also certain things that you're looking for. Search, machine learning and other things. There are also tools out there who can summarize contract. If you have one contract and you have 100amendments that change the contract, there are tools out there who can basically summarize that and show you a quote unquote non existing contract it created that has all the valid clauses that are active today. And that's a huge jump in productivity that will shift everything we do there. And I see similar advances in other areas as well. So for example, you could imagine that you have a data room and you basically ask the data room an intelligent assistant all the questions you have and it answers based on the information in the data room, what you do. So there's unlimited opportunity there. If you look at the automatability, you could basically look at the task and then see what are the different activities in this task. Look at each and every activity there and say is it automatable or not? Many decision support, but also decision tasks can be automated today. So that's a huge thing that in the industrial wave of factory automation we had 100 years ago, the work of the quote unquote normal worker was impacted. Now machine learning is more like a vertical automation progress that we have. So even highly sophisticated tasks could be easily supported, improved with modern technology.

Can you tell me about some of the tools that can help?

It's hard to get an overview right now, what tools are out there? So one thing that's definitely missing is transparency on which tools are available for which tasks and how do they compare. Today every practitioner would have to look at all the tools out there and then find his own decision matrix and then decision. That's a very tedious task. What we need, and you mentioned that, and you run this community, is an exchange of information, of experiences, just feedback from using things like contract analysis or analytics or natural language processing of documents in the data room and share it in the community. So in Germany we have a very active community with several working groups in specific topics regarding automation. Since about two months we also published some of the results, also in English language from the German association of M and A in our magazine, which is called M and A Review on the website mareview.com you could have a look there. It's a running discussion topic, publications, et cetera.

Yeah, that's actually really interesting how you have the network and you're sharing some of these ideas with M and A Science. The initiator was the fact that all these different companies have their own approach and I really focus on a lot of the people challenges. That seems to be some of the most difficult things to really solve. I don't think you can automate that stuff. I think there are certain things where we look at maybe more of the checkbox activities where hey, we're Looking for some pretty obvious things that are essentially the doomsday tomorrow type of risk. And hey, we got tons and tons of employment contracts, customer contracts, to really analyze and look for the variances which may identify other risks. The human to human part, we still have to do that. And I think that's been the area of focusing around the practices there.

But it's also certainly, let's say you have in due diligence, you have a question, something seems to be questionable. What's happening around a customer. Let's take a fantasy name like Bosch, for example, and you say, hey, this customer Bosch, what's cooking? And you talk to the person at the target company. Even in this situation, you could have support from a tool that shows all the relationships around this entity, Bosch, and all the massive information you have in the data room to give you a background without having to look through the 100,000 documents in the data room around this customer, Bosch and his complaints, for example, when drilling deeper into this topic. So clearly person to person is important, but it's also augmenting and assisting in using all the data in the data room that a single person is not able to capture as a whole.

Yeah, that's actually an interesting view. When you look at automation, what are the essential goals out of that?

The perfect automation would be basically two things. You have an end to end platform or data integration layer. You would have the opportunity to plug in different specific tools for specific tasks into that. And you're able to apply all the magic that we talked about, even in also coupling data from different phases in the M and A process, go back and forth and use the capability to not only compute, but also analyze and interpret large amounts of data.

So in short, you're basically cutting out a lot of the tactical crap so people can focus more on the high value activities that actually create value.

Absolutely.

Have you seen much interest in this or are people pretty satisfied with their current structure?

Yeah, there's a lot of interest in this. The key thing that keeps people in M and A and co op dev departments from rapidly adopting tools is it takes time to get a little more transparency, what's cooking in the market, and also some time to bulletproof that a solution is really the right one to use for screening or other activities in the M and A process. And there are lots of screening tools out there with lots of automation. But it's hard to figure out today without having time to try 10 different solutions. What's the right solution for you? And this is where we need to put in some Work and share experiences and information to have an easier time to select solutions which are right for the tasks that you want to automate.

I'm curious if maybe some of the most recent things you've worked on that you see some pretty early signs that you're hopeful around.

I think that we will see a widespread use of specialized end to end M and a project management platforms which are with or without a data room, but they basically run from a let's say portfolio management strategy screening all the way through the integration. Either the data room providers kind of fan out in the other phases or specialized project management system providers will provide that layer and would need to plug in data rooms and that's already there and that's a good starting point to further specialize automation tools like for example legal analysis of contracts in the data room. That's also already there. So a lot of good stuff that's interesting.

I think you got a good point where there's always that central repository of the information, but then the extension comes in forms of APIs to really be able to plug that in and centralize it. We're starting to see that. But then I think that also enables independency of companies to really focus on niche technology but then still bring it back into that central platform.

That's the idea basically to know what other data I need in different phases, for example in screening to compare different companies etc. And then apply everything on top of that based on a quote unquote standardized API model. So that's where we should be going hopefully in the near.

Why is it more challenging to integrate the businesses post M and A than the activities involved in the course of pre M and A?

That's a very good question and it's a widespread problem. I think that what you have to do is basically to focus as early as possible on questions and issues and problems and budgets and what have you regarding the integration. So at SAP we've massively changed what we do. I think in due diligence we focus a lot on how we can make the integration successful. So we look at the target, of course do all the target due diligence, but at the same time we also do preliminary integration planning and also trying to harden our plans to make them bulletproof, to make sure we have the budgets and resources of things we see coming up in post merch integration. And we do one thing which is we have the same people, for example in product due diligence and in the integration project regarding product aspect and that has increased the precision of due diligence results a lot because the same person who has to stand the challenges in the integration based on the analysis that you have done in the due diligence and that helped a lot to ease the pain that many companies see in integration. If you start early, if you have the same people doing that, that helps a lot to quickly improve what you do in due diligence to do better, better, better each time you acquire another.

Company makes sense really build this initiative early in the process. Does GDPR represent an issue with regard to the automation of due diligence or integration for sensitive personal or employee data?

Absolutely, yes. All the machine learning and analytics and even in a due diligence looking at a list of employees of the target all activities that are massively impacted by gdpr. And gdpr, if you look at machine learning is one of the key issues we have there because you want lots of data, but you can't use the data because it's personal identifiable information. You have to massively anonymize and do other things. So GDPR is an issue, but you also have to see that the PII information is only, I would say, a small part of the data lake. Let's say modern and sophisticated tools have ways to either avoid or anonymize these data.

Got it. You envision the tools you mentioned being available for small middle market practitioners or primarily only for large firms.

That's one thing that I was chasing. So at SAP, we were looking at different tools for screening companies and we saw that some of the tools are only offered as a service and some of the tools are either offered as a service or or software. We wanted to only look at things that we can use ourselves, not by hiring somebody for screening. And pay $50,000, give us some company names. I think there are still some obstacles there if you look at the screening tools. But there are also some new and innovative companies out there which offer more affordable price tags for small and medium sized companies as well.

Another question is there a best practice schema for ingesting a data room into a data lake so that analytics can be best performed? What tools have you tried so far which have given you the biggest impact on automation?

So I think the question is very good and it poses one important thing that we need for automation. We need a reference schema for the information or the data that we look at in the end to end m and a process. And one thing I've done in the information, my new book, I've basically built a quote unquote data model for The M and A process, what you see there, of course it's pretty complex, so that will be a part of the book as well. But to come back to your question, what I see is that there will be standardization of these data which make it easier to extract the information. I would also say that you don't have to necessarily extract the data from the data room to create your data lake. It might also be like a federated set of information from earlier phases data room later phases that you might access to create your quote unquote virtual data lake. So it doesn't necessarily have to be extracted. But for example, if you look at the complexity of data, if you put up a data model for the due diligence phase, I think I had around around 1000 data objects, about 1500 relationships. And that's only due diligence and not earlier in later phases. So we'll still take a little time to come up with an appropriate schema. But there are also smart tools which basically help you to get help and assistance to merge different data sources. So I wouldn't see that currently as too big a problem to move data from one place to another.

Your book coming out, can you tell me a little bit about that? I'm curious to know, for somebody that wants to learn more about this topic, what can they expect to get out of it? Is it high level stuff that kind of gives them an idea or you get a little bit more into some actionable things that they can actually apply?

Yeah, we wanted to basically create a book that's for practitioners. Besides the, I say more academic view of if a task is automatable or not, there are also a lot of very hands on things included there. For example, questions that you might raise when you run the task. If you do, for example, intellectual property, due diligence, what are the typical questions, what are the typical pieces of information that you're looking for, etc. So it has a lot of detailed information how to run the certain tasks, plus it has the information about if it's automatable or not. I'm also planning some further upcoming books which will have an overview of tools for these specific tasks as well.

How have you justified the cost of automating your M and a process and the proposed return on investment? In other words, what were the key areas in the business plan to justify the cost?

That's a very good question. I haven't done the computation yet, but I think the advantage you will get from having a lot of the information that you need to drive decisions MITIGATIONS maybe changes to the deal, more or less after very short time into diligence, will basically provide the value that pays basically for the increased speed in getting all the information and risks and things to be handled. And of course the value you get from the increased speed, but also the capability to process, you know, huge amounts of data. Finding paths Talked about the Bosch customer example. So what's happening around Bosch? It might be hidden in a thousand documents. So you don't have to spend another night reading a thousand documents, find something out and interviewing the people, etc. But you will have the information at hand and can basically follow the red tape.

Yeah, it's pretty cool. It's going to be an interesting world going forward. Do you have any advice you want to share in terms of others? It seems like we have some just general interest where people are curious about this and I think especially with a lot of corp dev teams that run M and A on large scale, where they want to be on the forefront of not only the people to people approaches, but also using the technology to be able to scale their M and A process. Maybe if you had some just general advice you'd like to share with those practitioners.

Yeah, maybe. Very simple, two things. So one is if you're interested to make the integration successful, you better focus on preparing to make the integration successful, not just signing the deal and leaving happily and somebody else has to take care of the aftermath. It's work that you have to put into plan for integration success. And the second thing is stay tuned to what's happening in the increasing order. Automation in the M and A process and the many things that we see in other places, be it Alexa from Amazon, you will soon have also on top of the data room to just ask questions, what information is in the data room, how things are connected, et cetera. The automation will be there, just grasp it and run the process better.

I agree. I like the Lexus example. I think that's starting to become a reality. I'm not sure if you saw the Google spreadsheets demo where they have a similar example where you can essentially ask a spreadsheet of data a question, it'll synthesize and give you a response back. Do you foresee the valuation of firms diminishing down due diligence to ongoing Covid crisis? If yes, do you see and recommend this to be the right time to ink the deals?

If you exclude companies in deep financial trouble? The first question is who wants to sell their company in the COVID crisis if they are financially viable? So it's not a market out there. We can say, oh now the strategists can easily buy cheap companies. If they don't want to sell to you because they want a higher valuation, then you have a tough time to cherry pick very good companies for a very low price. Again, unless they are in financial trouble and you come in as a savior. It really depends on the specific market for target companies that you're looking at. Even if the valuations, quote unquote in theory will go down because all the projections, et cetera, might be not very good with pre crisis numbers. It's really, I think, still a tough market because as a founder you don't want to sell if the market is down.

Have you seen more companies using automation during the COVID 19 shutdowns or has it slowed down?

Crisis, no matter if it's financial or health or others, is always creating pressure for companies to increase productivity to lower cost and it usually leads to an increase in automation. And that's also what we see today from an SAP point of view. SAP has always been sailing through crisis in a very healthy manner because crisis creates need for more automation. There's no question about that.

I would agree with. I think that's what I've heard quite a few times. I wonder what your thoughts are in automating your M and A process and how you may be able to do trending across multiple acquisitions.

So there are tools out there, if you look at for example companies like Midaxo or others that basically provide you with portfolio management and of potential deals in different pre deal phases and also analytics across different deals that will be done that are in due diligence, that are in integration. If you have like an end to end process management tool, there will be a lot of opportunities to compare different companies, different integration projects, different swim lanes in the pre deal phases for selecting different companies. I think that is already out there. I haven't looked at very detailed what exactly is there, but I've seen some glimpses of cross deal cross process analytics that it can use to compare improve things that you're doing in different deals.

Last question Dr. Carl, what is the craziest thing you've seen in M and A?

Yeah, I think one of the craziest thing is what I call cascading integrations. You basically acquire a company who has acquired another five to 10 companies and you want to integrate the company. But what happened in the past the other 10 companies weren't integrating so you suddenly have to integrate 11 companies instead of one. So that's certainly a huge challenge to integrate a lot more companies, entities, operations than you could foresee before you jump into due diligence.

That tends to lead to a pretty big mess.

Yeah.

Dr. Carl, I really appreciate your time today in this conversation. You had a book coming out. You want to sort of add a little color or where people can learn more about it.

Very simple. Go to my website drkalpop.com and there will be information about this.

Great. Hey, thank you very much. Thank you for taking the time to explore the world of M and A with our podcast. Please subscribe for more content and conversations with industry leaders. If you like our podcast, please support us by leaving a five star review and sharing it. I enjoy hearing feedback and connecting with our listeners. You can reach me by my email. It's Kisan K I S o n@dealroom.net M&A Science is sponsored by Dealroom, a project management solution for mergers and acquisitions. Additional educational content is available on Deal Room's blog@dealroom.net blog thank you again for listening to M and A Science. See you next time. Views and opinions expressed on M and A Science reflect only those individuals and do not reflect the views of any company or entity mentioned or affiliated with any individual. This podcast is purely educational and is not intended to serve as a basis for any investment or financial decisions.