“We Need a Context-Sensitive Controlling Approach”


Digitalization and startup culture are on everyone’s lips. Dr. Florian Heinemann, co-founder of the Venture Capital fund Project A, speaks about how traditionally positioned large-scale enterprises can learn from startups, venture capitalists and companies like Google and Facebook

Florian, you co-founded Project A together with Christian Weiss, Uwe Horstmann and Thies Sander. Which approach to startup financing do you use?

In the startup world, we make a basic distinction between consultancies, incubators, venture capital funds and business angels. Project A is a venture capital company (or VC). Like any other VC, we make use of funds in order to finance a portfolio of maybe 25 to 30 companies at a time. When an entrepreneur approaches us with an idea we try to get involved as early as possible, at the latest in the Series A financing stage. In Germany, typically one to five million euros are invested during this phase.

Which areas do you invest in primarily?

From a geographic point of view, we invest all throughout Europe. The first focal point here is on Germany, which makes up 40% of our investment, which is simply a result of the proximity factor. Our second focal point is around the capital cities, in particular London, Stockholm and Amsterdam, since we have the best networks there and the best relationships with co-investors. That’s important, since you’re always working in a network with others as a venture capitalist. No one is capable of investing something all alone. You get the best deal flow — meaning investment proposals — when you have a good network and you also possess a certain reputation in the market. The business is driven by reputation to an overwhelming extent.

And which topics does Project A focus on in particular?

We have a strong focus on the topic of digital infrastructure. No one can keep up with Zalando or Amazon today without having a very good technological foundation and a very solid infrastructure. Amazon can build its own infrastructure of course. 99.9 percent of companies

can’t, they lack the necessary know-how. They have to bring in third-party suppliers who give them access to the proper tools. We invest in third-party suppliers who have good ideas. As a rule though, we try to approach new topics every 18 to 24 months, because we are eager to learn. To succeed in this business, you have to be the fastest learner.

What do you aim to achieve with your investments? You coined the term operative venture capitalist for yourselves. What is that all about?

We would like to help startups to find the right configuration at an early stage. That’s our intention, the added value that we offer. We believe that the initial configuration of a company plays a very strong role in determining the path their development takes and we would like to take advantage of the opportunity inherent in this early phase to exercise a positive influence on the configuration of a startup that will make itself felt long into the future. That is exactly our idea of what defines an operatively supportive venture capitalist: we don’t just have the deep pockets, we also offer them support. Of course we only do that when an entrepreneur is interested in it. We employ 100 experts that advise our companies operatively and pitch in themselves. We don’t receive any further shares for providing this support — instead we just bill them for the costs of services rendered without trying to make a profit on top.

Dr. Utz Schäffer (photo: Michael Jordan)

Why do you consider this support aspect to be so important?

When the technological foundation in particular isn’t right in the beginning, then you have to build it from scratch all over again after three or four years of growth. Especially in the area of core applications, we can see really clearly how useful operative support at an early point in time can be. If we already tackle the configuration of a company in the right way from the start — the organisation, systems, processes — we open up an opportunity for the company in question to set itself on another development trajectory.

“To succeed in this business, you have to be the fastest learner.”

With your mention of processes and systems we have also already come to a central topic for us, controlling. Does controlling play a large role in startups?

(laughs) It plays a central role for the good companies in the digital sector, as such companies are steered to a great extent through data and KPIs. You can really say that these companies are also successful with this approach. If controlling is systematically good here, it is not perceived as the area that gets on everybody’s nerves — instead it’s seen as a real part of the value chain.

I have the impression that controlling in startup contexts is frequently reduced to KPIs. The linking of the KPIs to strategy is missing.

I think it depends an awful lot on which stage the startup is currently in. In the early stage, the financial area is usually limited to managing shortfalls or bottlenecks, since the strategic objective at this point in time is super clear: a startup has to demonstrate three to five developments to the investor in order to receive the next round of funding. That’s why founders don’t really think much at all about strategic issues in the beginning. I think the phrase “driving by sight” captures that really well. Especially in an early phase a startup has to prove its structural reason for being. Either it has to be able to keep a story alive, a fantasy, like a Research Gate for instance, or it has to be able to buy or acquire customers through expenditures that could pay off at some point, like with Zalando. You also have to consider that a lot of startups also need five, six or seven years until they see some kind of payback.

Shouldn’t that make a controller nervous?

In the beginning, we just don’t know which startups are going to be successful and which aren’t. One thing is totally clear though: we have our 25 to 30 investments in one fund and know from the start that 80 to 90 percent of the returns are going to come from 10 to 20 percent of the investments. It’s like that for every fund. The really good venture capital managers don’t have a lower failure rate, they actually have an even higher one. The more successful players in the venture capital sector fail more frequently, because they take greater risks. On top of that, it also takes time before a company becomes successful.

What would you recommend to large companies: that they put everything together on their own internally or that they would do better to purchase already existing startups?

There isn’t any one single approach here. I think that you always have to do different things. The good startups are not going to go to the incubator of a corporate partner. The most important thing for a corporate player to do is to build up digital competency, internally as well. Large companies have to struggle with a massive selection problem on all levels. That’s why it doesn’t work without serious internal digitalization efforts or without a central, competent digital M&A team. In my opinion, a commitment to building such a team over time is of extreme importance.

Dr. Florian Heinemann (photo: Michael Jordan)

Can you perhaps give us some examples?

The Axel Springer SE followed this path with great success and then started to invest in a relatively large amount of participations through a central team. It can also make sense to work together with venture capital funds like us, but that can be an activity to round out your portfolio at the most, in order to get a couple new impulses. The Otto Group decided to do that for example and they invested with us. I would really recommend that everyone take this centrally managed approach, as well as that they take a look at the M&A teams at Zalando, Facebook and Co. On the other hand, I think that decentralisation, as practiced by Volkswagen for instance, is a big mistake, since it is extremely hard work to build up a competent digital M&A team in every business area. I think the chances of pulling off that approach are rather low.

What makes Google and Facebook so successful in this area? What can companies learn by studying them?

Google and Facebook very consciously buy things that belong to their core business and expand upon it. The focus of their M&A activities is not on buying business models though. Instead, they invest in competences — since you can make a business model your own quickly, but that’s not necessarily the case for new competences. Facebook for instance decided for itself that virtual reality capability in the area of social interaction is probably going to become important at some point, so they bought Oculus. Springer has also demonstrated good foresight and invested in intelligent classified advertising models. You have to be able to recognize those kinds of developments though. And you also need a CEO that can take the hit of a failure now and then.

“It doesn’t work without serious internal digitalization efforts or without a central, competent digital M&A team.”

Corporate players like Daimler and Co. are positioned in a totally different way than Google or Facebook.

Yes, I think that’s precisely the key to the whole thing. In order to be able to do it well, it really helps to have done M&A before and also to have produced digital lighthouse projects in the past. That’s why I absolutely recommend that companies launch a variety of M&A activities, while at the same time developing their own idea pipelines, even though that may be difficult to implement. You always need both! Taking a look at pharmaceutical companies may help to illustrate my point better. For these companies it’s absolutely clear: a pipeline that is only driven by M&A, one where whatever active ingredient startups are bought up and integrated into the company’s own distribution structure, such a pipeline is definitely more expensive and worse from the point of view of long-term return on investment than developing their own. Especially in an M&A market as transparent as that of the pharmaceutical industry you have to have an idea pipeline too. The digital sector is a similarly transparent market.

Can you give us a couple examples there too?

The Axel Springer SE would be capable of producing its own digital projects now for instance. These days they have a lot of M&A experience there and they’ve seen that the digital business doesn’t just burn up money, but that it also earns money too. The CEO would have the credibility now to do that. Family-owned companies are in a good position to do that too. The Viessmann family for instance has set its sights on just that and founded an internal incubator in Berlin, where they’re developing new digital business models around the topic of smart homes. Otto is applying serious pressure, constantly starting new initiatives with new shops, sales concepts and so forth. You have to really make the commitment.

Do all companies have to reinvent themselves digitally?

I consider this approach to be the only real choice, as there’s no doubt that digitalization is going to change the value chain. If a company does nothing, it’s definitely going to lose its position or drop to the wayside altogether. I believe in a mixture of direct investment, venture capital investment, an internal innovation pipeline and digital transformation of the core business. However, the external initiatives are the more important part here, even if they typically appear small in regards to total sales volume. That’s the misleading thing! You have to realize though that in all likelihood it is precisely there that the state-of-the-art digitalization know-how will take shape, which will then be able to further help with digital transformation.

That means that companies also have to learn to think differently. A new mindset is required.

Exactly. It is really crucial to realize that you need both: you need business, but most of all you need competence. The central question is surely: How can I build up the vertically specific digital competence, that is, branch-specific business models or value chains? The activities have to be aligned to do that. This way of thinking has made Google as wildly successful as it is and explains for example why Google bought Deepmind in 2014, a startup specialised in programming artificial intelligence. True, Deepmind doesn’t generate any revenue, but it is a relevant competence building block for Google.

How can you actually determine the value of such a competence building block?

Controllers really do indeed have to develop an understanding for the fact that competences also have a really great value, maybe even the more important one. They will have to think about how they can process, assess and advocate this internally. I think that is extremely difficult for many of them. This is another instance where you can learn a lot from pharmaceutical companies, which of course know exactly that the patent for medicine X is going to expire in 15 to 20 years. Then the generic brand manufacturers will come along and within a very short time the margin will disappear from the model. You actually know that for all other business models too. Every business model has an end date. You know that at some point it will lose margin and the capacity to function correctly, even if it doesn’t happen that abruptly. In the startup world it is exactly like it is with a pharmaceutical company, only not as obvious, and as a result not everyone realizes it. This way of thinking, this way of coming to terms with the periodic nature of things, is enormously helpful though!

… and that is not really a part of the traditional controller mentality.

Correct. It’s difficult for controlling to assess the value of competences, since their impact frequently lies very far away or it is very indirect. That really does require a certain kind of hypothesis-driven approach, which is not easy of course in a world with limited budgets. That’s why I believe that we need a context-sensitive controlling approach here, to deal with things. We have to learn to understand the startup world and to be able to say: the startup in question is currently in this or that stage and these or those indicators are important at this point. The controllers have to really get it that things like EBIT are more or less unimportant in a very early stage. The first thing on the agenda is to move towards structural viability — in the sense of “can this model pay off at some point?” If the startup generates decent customer value, if the customers like it, then they will come back too. Because precisely that is the prerequisite for being able to create a functioning business at some point.

What does that mean for controllers exactly?

We need controllers that are capable of saying “now here I’ve got case A, so I need to apply this model and maybe just present two pages with the highlights instead of also submitting the whole 47-page-long reporting form”. When the startup has developed further, then case B comes into play. What’s necessary is a sort of stage-model for every type of business. Controlling departments definitely have to learn to do that, and M&A departments do too, just as much.

“You can make a business model your own quickly, but that’s not necessarily the case for new competences.”

Should controllers also push more insistently for thinking in cash flows in that case?

I don’t see the need to focus so much on cash, especially not in the early startup stage. I would call it more of a monetization potential focus, in the sense of “can that make money?” “Is this something that customers are likely to be willing to pay for?” You have to be careful though, because it’s a very fine line here. Often you’ll hear entrepreneurs say that none of the venture capitalists in Germany are visionaries. Then you’ll hear them say: “I’m too early for my idea, I’m too early for the market.” Whereas in reality the idea is just a bad one!

I would probably have bought that argument totally naively for a lot of cases…

I can’t just buy into it without questioning it. Facebook wouldn’t have worked a couple of years earlier for different reasons and wouldn’t have been a good idea either as a result. I think you do have to take a really critical look at that stuff. Otherwise it would be an open invitation to say “It’s not focused on figures” and “carte blanche for everything”. I consider defined objectives and deadlines, this balanced scorecard way of thinking, very helpful in this case in particular. For me, those are the micro-conversions. The ultimate conversion, that’s the purchase. But in order to get there you have to achieve a series of things. As such, we attempt in principle to introduce a systematic approach to something that is less measureable in a hard sense, less hard in the sense of “less monetary”. We consider that to be of the utmost importance. It’s also precisely here that the assurance of rationality that characterises controlling comes into effect, which you guys at Vallendar always propagated very strongly.

Rationality is context-sensitive to a high degree in this case.

Yes, that’s why we need controllers that are willing to investigate every individual case that are then able to say: “In your case it looks like this or that.” They have to be capable of modelling that and not just resorting to formula F — without becoming arbitrary or allowing arbitrariness in. That’s a relatively creative job. I think you need both. You have to be able to separate them consciously though and say: “We’re going to do this here — and we’re not going to make any compromises in applying it strictly.” In certain areas there’s no degree of freedom and no excuses either. The cash flow has to be right. Google does that with its Adwords business for example. That’s managed in great detail with no compromises and the sales targets are clearly defined. At the same time though you’ve got to create islands that have a higher degree of freedom too.

Apparently that is almost always problematic…

Yeah, people have an extremely hard time with the parallel nature of different controlling contexts.

Last but not least: What does the optimal controller of the future look like?

We probably need more controllers that come from the new business or innovation sectors, since they bring the right mindset with them from there. I have an image that I really enjoy using: “The new right is the less wrong.” The tolerance for mistakes, the acceptance of uncertainty, is important. A lot of things often don’t happen, because people say: “I don’t have the perfect approach for that yet, so I won’t even try at all.” In an agile enterprise or in a startup it’s not at all about reaching 100 or 95 percent, it’s about first going from 40 to 60 instead. That’s a totally legitimate process, and we need people who can live with precisely that. That’s extremely tough mentally for a lot of people, that: “Being less wrong also means making progress.”

And who could fill the bill best?

As I said, in my opinion people from the new business or innovation sectors would be more capable of doing that. In any case, that requires a very high capacity for abstraction though — similar to marketing, where a broad spectrum has to be modelled as well, from “Does this TV ad trigger emotions?” all the way to bid management in Adwords for 20 million keywords on the basis of lifetime values for 180 days in consideration of the contribution margin etc. Even a chief marketing officer can frequently have an extremely hard time with this ambivalence. The question is basically: “Which person works best here?” My observation in this case has been that people who come from the quantitatively oriented side of things but possess mental flexibility, openness and a capacity for abstraction manage better to develop themselves there than individuals from the more qualitatively oriented areas. I think that is an easier development trajectory than the other way around.

In principle that speaks for drawing on the existing controller profile…

I think that’s a better way than starting to bring too crazy people into controlling. But you do really need both. Even though mixed DNA is always tough, as much as specialist departments with indefinite DNA: you need someone who is capable of managing both poles.

Florian, thanks a lot for the conversation!


This interview first appeared on Controlling & Management Review (61/6, 2017) and was conducted by Prof. Dr. Utz Schäffer, Director of the Institute for Management and Controlling (IMC) at the WHU — Otto Beisheim School of Management in Vallendar and co-editor of Controlling & Management Review.

Photo credit: Michael Jordan.

More stories about