Striking a balance in R&D portfolio management

How can R&D managers deal with the conflicting requirements of running a portfolio of projects for best short and long-term performance? Delegates at this year’s annual conference explored how to build a balanced R&D portfolio in a world that is changing increasingly quickly and yet faces issues that demand long-term solutions.

The job of an R&D manager is increasingly like that of an investment banker. Both bet on the future worth of a portfolio of investments. Both try to balance small and large bets. Both try to hedge their bets, so that the failure of one will not destroy the portfolio. And both are expected to deliver great performance over both the short and long term. So what’s the best way of balancing these conflicting criteria?

According to Walter Steinlin, head of the Innovation Competence Center of Swisscom, the challenges of building a balanced R&D portfolio are much as they have always been. How do you manage short- and long-term projects at once? Should you make a few big bets, or many smaller ones? Should you focus on evolving what you have, or look for breakthroughs? Who contributes what in a partnership? And how do you predict whether a nascent project will become a route to a long-term business or simply solve today’s problem?

The idealised R&D portfolio is a series of carefully chosen projects flowing smoothly through a project pipeline, checked at intervals by gating functions that consider their worth and suitability for the business units that might apply them.

“But really it’s more of a Brownian motion than a flow, and disruptive projects can become a Brownian target, which starts to shake your portfolio,” said Steinlin. “The old Stage-Gate^® process continues, but it’s increasingly volatile and iterative.”

What’s changed is that today’s innovation processes have more dimensions than a simple linear pipeline. New processes can mean greater efficiency. Redesigning the customer experience can improve today’s offerings. Breakthroughs can lead to new products and services. Venturing strategies may launch new companies. But what investments do you make in each of these approaches?

There are contextual issues to deal with as well, especially for those in the telecoms industry such as Steinlin. The rise of the worldwide web has made many processes much more open. Steinlin is now wondering whether he should be drawing on ‘the wisdom of crowds’, as a recent best-seller described the insights that may be available through web-based collaboration mechanisms, as part of his innovation process.

“The challenge is how much do we have to open up and do we really have to partner with everybody?” he said.

Another issue is that many of the products in use in the telecoms industry are now in a permanent beta phase of development, never quite being finalised. He compares the release history of Microsoft operating systems, which come out at intervals of years, with that of the Flickr photo sharing web site, where the functionality is updated monthly or even weekly.

“As an R&D manager does that mean I can ever get rid of a project, or do I need to plan for permanent projects in my portfolio?” he asked.

Long-term portfolios

Massimo Comparini, industrial integration director for Thales Alenia Space, has a different set of portfolio management challenges. His business is affected by geopolitical issues that can lead to dual-use investments in space equipment, and emerging nations buying satellite systems to promote their regional leadership. Meanwhile the industrial context for space engineering is changing, as competition increases, technology life cycles accelerate, and countries such as Brazil, India and China take a growing role.

Space technologists also have to deal with an extremely broad value chain. Comparini said that a constellation of satellites, for example for earth observation, may cost billions of euros. Individual satellites cost tens or hundreds of millions of euros, while their subsystems will cost between €2m and €20m, and the underlying technologies a fraction of a million euros apiece.

“This broad range cannot be sustained by any one company,” he said. So space technology companies need to understand how application areas such as telecommunications or earth observation will evolve. This enables them to understand the system trends that define the product requirements that, in turn, act as inputs to the product portfolio and the R&D process.

This prediction job is particularly critical for the space industry, because of the extreme difficulty of servicing its products. A telecommunication satellite has a service life of up to 17 years, but communications technologies evolve much more rapidly so flexibility is vital. This means developing a variety of system architectures to enable that long-term flexibility. In turn this means working with two plans: a short-term plan that looks three years ahead and defines developments based on a product portfolio analysis; and a longer-term plan, looking up to 10 years ahead, which considers what technology breakthroughs are necessary to keep meeting customer needs.

“The question then becomes how you balance the extra cost of creating platforms for long-term flexibility with the necessary shift to standardisation that will come as technologies evolve,” said Comparini. He concluded that a balanced R&D portfolio that extends beyond company borders is vital to keeping the proper balance between product portfolio evolution in the short term and the longer-term outlook.

“Co-operation with a large network of external R&D centres is the only way to sustain a long-term vision and make it compatible with shareholders’ expectations of profitability,” he added.

The Google way

By contrast, Google is moving so fast that its portfolio appears to go largely unmanaged. But that does not mean the company does not have a long-term strategy: its stated mission is to create democratic access to information.

“The best search engine should give you exactly what you want, whenever and wherever you are,” said Robin Williamson, engineering director of Google’s Zurich engineering centre.

Google is building a rigorously equal research model, which uses web technologies to support and share engineering efforts around the globe.

“More than 50% of our users are outside the US and more than 90% of our users don’t use English as a first language,” said Williamson. “Searches are also contextual: to the French, Cote d’Or is a region, whereas to the British it’s a chocolate bar. So we set out to create a global engineering organisation that helps us understand local markets, find talent and localise our products.

“We don’t do traditional product management, thinking small or going it alone,” he added. “All information and visions must be shared across all the teams. We focus on the user first. The money will follow.”

Google applies a number of rules to its headlong innovation strategy. The first is to emphasise the ‘dos’, making it plain what is the preferred behaviour. The second is to create that global engineering organisation in which everyone gets to contribute on an equal footing. The third is to hire the best, using peer group interviews.

“What this does is help us identify candidates with the smarts to deliver for the company and the ability to think in different ways.”

Google also encourages engineers to be entrepreneurial about their ideas, for example by allowing employees to spend 20% of their time on their own projects. There are some more prosaic innovation strategies at Google, too: ‘fix-it groups’, which might be called tiger teams elsewhere, focus on solving one issue; ideas databases and solutions request boards; and ‘hackathons’, in which teams of staff go offsite for a few days to uncover what might become longer-term projects.

Transparency is another part of the strategy. Williamson says that Google engineers have access to most of the information in the company (apart from the core search algorithms), so they can find out what everyone else is working on.

Google also focuses its innovation on the user: if you lose the user you lose it all. And it thinks that speed matters. Williamson quoted Rupert Murdoch, head of News Corporation, as saying: “Big will not beat small anymore. It will be fast beating slow.”

So what makes this apparently free-flowing environment work for one of the world’s most lucrative advertising brokerages?

“There is something very fundamental in Google that allows us not to go through that planning process,” said Williamson. “Part of it is the company, and from a technology point of view there are almost unlimited resources to build software and test it, from the very simple to the very grandest scale. There’s also natural selection by peer group of where our engineering effort is best focused.”

Welcome to the Bazaar

Google appears to be managing its research portfolio by not managing it. Instead it relies on a combination of a clearly articulated and easily understood long-term vision; a series of innovation principles that emphasise speed, transparency and peer review; and a community of engineers who are expected to act as entrepreneurs and promote their wares in an open ideas market.

It’s easier to take such an approach when you are starting from a clean sheet of paper, as Google did, with highly motivated and capable staff unburdened by any legacy business. But other organisations are evolving their innovation strategies in similar directions, enabling their portfolio management efforts to become more a function of their structure than an explicit discipline of its own.

Paul Friedel of the R&D group at Orange Labs, part of France Telecom, explained how the old, layered model of the telecommunications industry is collapsing as companies try to move into each other’s sectors for scale, consolidation and greater profitability, and as customers’ views on what they can expect for free changes. France Telecom does have some advantages here, since it has 160 millions customers worldwide and knows how to service them, in terms of handling maintenance and billing issues. But it’s not (yet) like Yahoo! or Google, providing services which are free to the end user.

In this context, a portfolio management strategy is perhaps less valuable than new attitudes about what you’re trying to achieve.

“We need a new way of looking at R&D as one branch of our innovation engine,” said Friedel. “We believe there is not one good model for innovation but that we need multiple innovation models.”

The result of this thinking has been the emergence of the Orange Innovation Bazaar, a group of organisations, ranging from the R&D function through marketing and trend trackers to the national business units, which work together to innovate in a variety of ways. For example, in the Exploration Centre, new product and service ideas are designed with customers to see if they have market appeal. The TechnoCenter, on the other hand, drives the development process forward.

“To live inside this approach is quite exciting,” said Friedel.

The company is also trying new innovation approaches. Its 3Partner projects link R&D, marketing and implementation teams to achieve the kind of time to market needed in the fast-moving telecommunications industry. The Exploration Centre links a new business model idea with the R&D and design functions, informed by focus group studies, to create new service concepts. To keep up the pace of innovation, its work is taken to market three times a year in spring, summer and autumn ‘collections’, an idea borrowed from the fashion industry.

Orange Labs is also wrestling with the increasing importance of the customer in their business. Where once customers were passive consumers, “customers have become users, and reflecting that we understand that we have to serve people trafficking on our networks [for free] almost as well as we serve our direct customers.

“What we are really moving to is putting the user into the centre and really trying to think of the best way to serve that person, because these people also create [in open source projects], act [in social networks], and even administer their own networks.”

So the grand challenges for Orange Labs as it tries to make its R&D effective include making the Innovation Bazaar work, making money out of new services, and creating trust with the users. It also needs to offer users tools so they can protect their privacy at will, bridge the gap from abstract academic concepts to pragmatic usage, and grow its resources at the right pace, so that the burden of enabling internet activities doesn’t lie disproportionately with the network providers.

This is a lot to ask, especially of a research group that used to be part of a state-owned monopoly. But it is clearly seen as vital for France Telecom’s ability to innovate over the long term.

“The whole management structure is pushing for us to have a much closer relationship with marketing,” said Friedel. “What I’m trying to do is create the link with marketing so we are not simply doing what they would like us to, but so that they are learning from us as well. I want to be completely free, but still have a good dialogue with all the stakeholders in the business.”

Collaboration at scale

Two major efforts to avoid missed opportunities at the industry level, through multinational co-ordination and collaboration, were also discussed at the conference. But it is unclear whether the effort and delay involved in creating the co-ordinating structures will be worth the gains from the co-ordinated effort.

We’ve covered the Innovative Medicines Initiative (IMI) in a previous issue of eIQ. The idea is for Europe’s major pharmaceutical companies to work with each other and external groups, such as patients and doctors, to find better ways of developing drugs. The work is pre-competitive - it’s about tools rather than products - and has been widely supported by the Europe pharmaceuticals industry.

But creating the IMI has taken four years, a substantial proportion of the average amount of time (10 years) that it takes for a new drug to get to market.

“Why did it take so long to create the IMI?” asked Karima Boubekeur, who is head of external research and innovation environments for Roche and has taken a leading role in the development of the IMI. “Part of it comes from the legal and political constraints of being a community body within the European Union. And there is a cultural difference between industry and the Commission, of accountability versus responsibility, and decision versus consultation.”

Boubekeur says that her strategy for getting through the paperwork was to study the European Union’s legal framework, negotiate and be patient, trying to understand the different constraints on Commission counterparts. But it was still a lengthy process.

“It took two years to do the content and two years to do the politics,” she said.

Jan van den Biesen, vice president of public R&D programmes at Philips Research, had a similar experience in his role as chair of the Artemis project’s working group on governance, funding strategy and Joint Technology Initiative structure.

Artemis is an admirable idea to create a three-way partnership between industry, academia and the Commission to sustain Europe’s strong position in embedded systems, the computers we never see but that run everything from washing machines to aircraft engines. The initiative was launched in 2004 by European industry and the Commission and immediately attracted 10 out of Europe’s top 25 companies by global R&D spend. It now has participants from 24 countries, as well as representatives from SME federations and earlier relevant initiatives such as ITEA2 (on software issues), and MEDEA+ (on hardware issues).

According to van den Biesen, one of the aims of Artemis was to pull together the previous, rather fragmented efforts into a strategic research agenda backed by a public private partnership with both critical mass and a dedicated legal structure under Article 171 of the Treaty of Rome.

“For all involved, this was terra incognita,” he said. “What we had in mind with the Artemis joint technology initiative was to create a European Research Area for the topic.”

There followed a four-year negotiation with the Commission to formulate the Joint Undertaking in a way that met EU rules and provided the co-ordination and incentives necessary to make Artemis worthwhile. The project has ended up with a total budget of €2.6bn, including €410m of European community money that is being used as a top-up mechanism to encourage funding from national governments. Again, though, the time that it has taken to formulate this collaboration structure doesn’t fit well with the pace of the industry it is supposed to promote. In four years, the performance of the microchips at the heart of all embedded systems quadruples.

van den Biesen hopes that the work he and his colleagues have done to create Artemis will become a toolbox that others can use to shape the European Research Area. But he highlights a mismatch between the motivations of the Commission and of industry.

“Our feeling is that the current institutional system is dominated by risk avoidance rather than high trust,” he said. “Everyone is just waiting for someone to make a mistake. If we want to change this we should try to adapt financial regulation to take into account the realities and risks of R&D funding in global markets.”

The end of planning?

So should R&D managers continue to try and plan a balanced portfolio, matching sectoral risks, the long and short term, the speculative and the evolutionary, in their choice of projects? Or should they regard their role as being a coordinator who, through clear leadership and communication, sets a context in which networked innovation flourishes yet always moves towards to meeting their company’s goals?

If it’s the second option, how should R&D managers handle what will become their new portfolios, of stakeholder partnerships?

One of the first conclusions from the conference was that you can’t expect to run successful external collaborations if you’re not collaborating well internally first. Since successful R&D is becoming just part of successful innovation, R&D managers need to draw in senior management, marketing and other disciplines who can enable and contribute to innovation. And since good people are hard to find and harder to keep it may be time to start thinking about employee satisfaction as much as customer satisfaction as a key part of long-term portfolio management.

Making an organisation collaborate well takes a number of steps, such as building internal trust in complex environments, implementing the supporting tools, organising data sharing and training the staff. Creating a benign intellectual property framework helps, and trying to co-ordinate your internal organisation globally before entering global collaborations with third parties is also an important step. The next is desire.

“These kinds of things can only work if you have people who want to work together,” said Boubekeur. “In the case of the IMI it was the European pharmaceutical industry R&D directors who first came together.”

Trust and dialogue matter too, as does ensuring that partnerships are structured in a way that is appropriate to their size. Steinlin said that in partnerships with over five members, you need to do some serious preparation.

“But with small partnerships you can just do it. Everyone knows what they’re doing and everyone knows what is expected of them.”

Monica Schofield of TuTech Innovation highlighted the perils of ‘unequal dependency’ in partnerships, that is the mismatch between the various parties’ needs: for example, academics who need to work with long timescales while small to medium-sized enterprises need more immediate returns.

There were also questions about the balance of value in very large collaborations. Although it takes a lot of time and effort to create Joint Technology Initiatives and Framework Programmes, once they’re formulated they make collaboration much easier.

Ken Taylor of the European Technology Network warned that public/private partnerships, no matter their size, tend to rely on a standard structure that creates attenuating interfaces that slow everything down.

“You need to focus on the project outcome, rather than exhausting yourself on creating the project,” he said.

Boubekeur, having just been through the creation of the IMI, agreed that it is the outcome, not the process, that is key. But she questioned how you could bind people whose concerns centre on running a process to the success of the outcome. She also defended the value of the Commission’s role in formulating the IMI.

“The European Commission played a catalytic role in giving us the idea and in bringing other stakeholders to the table with a totally different mindset. It really made a change in the way we work together and made us think differently about partnerships and collaboration.”

van den Biesen pointed out that ‘related variety’, the concept of creating an ecosystem of partners with complementary skills is a good idea.

“If you collaborate with partners that are too similar to yourself you don’t learn much, and if they’re too different you can’t work well together. There is an optimum that you get better at reaching as you learn more about partnering and bridging the gaps.”