Leader.co.za - Management, Training and Career Advice for Business Leaders


Telephone: 1-650 851-2095
E-mail: Send us an e-mail
Web-site: Visit our web-site
Postal Address: PO Box 620955, Woodside CA, 94062
15 APRIL 2011
Getting the right people on the right projects

by Julian Birkinshaw: Professor of Strategic and International Management, Senior Fellow of the Advanced Institute of Management Research and Deputy Dean for Programmes in London Business School.

Matching project needs with individuals’ preferences can be tricky, says Julian Birkinshaw. He shares the story behind the development of an innovative system at AdNovum designed to do just that.

How do you get the right people on the right projects? This is a perennial concern for anyone running a large consultancy, an IT services company or, indeed, any sort of project-based organisation. Most companies default to a top-down allocation system: people are put on projects according to availability, favouritism or pure luck. It is reasonably efficient, but it gives little consideration to the needs and preferences of the employees being staffed. As a result, some lucky folks get all the exciting projects; some get stuck on a project-from-hell forever; and some never get a chance to prove themselves. Little wonder that turnover in these types of companies is often 20 per cent or more per year.

What are the alternatives?

Some companies, such as Eden McCallum, the London-based strategy consultancy, have experimented with a virtual-company model in which they act as the brokers between freelance project staff on one side and clients on the other. It’s a clever approach, but it provides no guarantee of work for the freelancers; and it is hard to scale, because it relies on the team in the middle making a high-quality match between freelancer and client.

Another option is often called the ‘internal market’, sitting somewhere between the traditional and the virtual models. On paper, the internal market sounds attractive: it combines the efficiency of the top-down model with the fluidity and sensitivity to employee needs of the virtual model. But in practice, well, it turns out that there are very few examples of companies using an internal market system on a consistent basis. This is why AdNovum’s experiences are so fascinating. This Swiss software company started out with a highly unusual approach to project planning; but, little by little, it evolved, emphasising some parts and downplaying others. Today, its systems are ostensibly similar to those of many other software companies, but it still holds on to a few key differentiating principles that help it retain its vibrancy and speed of movement in a brutally competitive market.

An established player

Founded by entrepreneur Stefan Arn in Zurich in 1998, AdNovum is an established player in the provision of security and software services to clients in the financial, logistics and public sectors. The company has a full-time staff of 240 people. Its business model, according to Arn, is about “hiring highly skilled and ambitious people, giving them interesting, tricky, work to do and delivering software services of the highest quality.”

For the first decade or so, AdNovum was a small software provider, living on its wits and with no formal systems at all. Some innovative solutions for its biggest client, UBS, put the company on the map; and it began to expand rapidly. By 2002, Arn realised he had to start thinking carefully about such things as staffing and project planning because, as he put it, “I wasn’t able anymore to schedule all my people on the projects. It just became too complex, keeping all my employees’ skill sets on the top of my head.”

Arn wanted to create a flexible system in which people could be moved quickly from one project to another as well as work on projects that excited them. “I had seen a small company manufacturing parts for cars. They standardised their approach as much as possible in terms of manufacturing, so that they could very dynamically allocate their people to whatever they were required to manufacture. And that was my main source of inspiration: I realised I could improve my planning process but first must completely standardise my software manufacturing.

“So we developed a Facebook-like system on the Web where you as an employee could portray yourself, especially your skill set, for others to review; and we encouraged people to keep their CVs and their skill sets as accurate as possible.”

Armed with this wealth of information about employees, Arn figured that project managers would then be able to select the best people for each piece of work. But, of course, it wasn’t that simple. Very quickly, the new system became something of a competition: people could apply for a project by looking at the pipeline of new projects and deciding which ones suited their skills and preferences, while at the same time project heads were calling up the people they wanted, based on their profile and their track record. As Arn explains, “That led to a situation like in school classes, when you try to split the class in two to have a soccer match. It was always the same people being picked first; and it was always the same people at the end, waiting to be picked. This made it possible for me to develop a ranking system. I counted how many times certain developers were asked for by projects, and out of this I got a performance-based ranking.” Rather than keep this information to himself, Arn posted it on the Intranet for all to see. He also linked the individual’s position in the ranking system with their bonus. “What then emerged was a performance-based culture, with people striving to be the best.”

While this market-like system fuelled the competitive instincts of the AdNovum employees, it still needed to be squared with the practical realities that some projects were more important or on tighter deadlines than others. So Arn put in place a weekly staffing meeting called Cluedo, after the famous board game, to match real delivery dates and contractual obligations with the allocations made through the internal-market system. Cluedo meetings were open to all employees, with six to 15 people typically attending; and they were deliberately short – 60 minutes maximum. Recalls COO Christian Crowden, “We wanted to have this open and participative approach. It was important to make sure that if people did not agree or didn’t understand why they were in a project or they were reallocated, they could always come up and ask.”

How well did this system work?

AdNovum grew strongly during the noughties and gained a strong reputation for responsiveness, at least in part due to their innovative staffing model. As Arn comments: “When I’m reading the software engineering literature, they’re always missing the most important point, which is that you get an agile workforce when dependency on a single individual is minimised. We were super-agile in allocating people to changing project demands. Let’s say we ran into a problem with a release delivery, we were sometimes able, overnight, to double the workforce working on a project.”

Evolving the model

But the story doesn’t stop here. AdNovum continued to grow; and in 2007 Arn sold the company to IHAG Holding AG, moving on to other challenges. Leadership of the company was handed over to Ruedi Wipf (CEO), Crowden (COO) and four colleagues, all of whom had been with the company for a long time.

From a team of 45 people in 2002, the company now has 240 people; and, in addition to the main office in Zurich, there are 60 people in Budapest, 16 in Bern and four in Singapore. In a typical year there are 60–70 customer projects, as well as 20–25 internal projects running at the same time, roughly a fourfold increase since 2002. So, in all respects, the staffing challenges facing the company have changed considerably and the model has moved on accordingly. Crowden reflects on the years since Arn’s departure: “The model had to evolve. It’s interesting: the company has now grown considerably, and it doesn’t work anymore in that flat structure where everybody applies for everything.”

How has the project planning system changed?

The principles are the same, but the day-to-day processes are somewhat less market-like, less openly competitive. It has also become more structured, with sophisticated software tools to make things run more smoothly.

The focal point is still the Cluedo meeting, in which the objective is to ensure the best possible staffing of project teams based on the resources available. The provisional plans are made using software tools, with 10 project managers putting together the day-to-day schedule across 60 or so projects. Then, in the meeting they discuss the subjective parts – who is working with whom, who has asked to be moved. “We only look at the problem cases, which obviously tend to be projects in their early stages or the closing down stages, as well as those that have bumped into problems along the way,” says Crowden.

The Facebook-like system has, however, lost its importance and has been replaced by a more informal process, essentially built around the project managers’ strong personal knowledge of their teams’ motivations and skills.

Two years ago, an existing component was linked to the planning process that tracks actual work being done – essentially a real-time time sheet showing how programmers actually spent their time in the previous days or weeks. “We learned over time that we were sometimes in our happy little world, doing our planning thing, when in reality the guys were working on something else,” explains Crowden. “So this system links actual and forecast hours nicely.” This system also ensured buy-in from the programmers. Every day, when they logged on, there was an automatic view of their personal allocation. “We wanted to make sure somebody would realise the plan was going awry and say, ‘Hey, but I am actually doing something completely different; why am I planned this way?’”

So how does the new way of working take into account the needs of programmers?

The Facebook-like system, in which programmers describe their own skill sets, has lost its importance, perhaps because of the high level of tacit understanding among project leaders. Says Crowden: “There is still a lot of informal communication, with everyone in the company represented by someone in the weekly Cluedo meeting. So if someone says, ‘I like to have two or three projects at the same time, otherwise I get bored’, then that can be arranged. Or there are other software developers who need to be able to change the context now and again, because they just say, ‘Okay, now I’ve done this for so long and I want to work two hours on some other project.’” The company also has a formal policy of rotating people, perhaps once every 6–12 months for a junior person and three-to-five months for a more senior one. “This helps the self-actualisation of people, it keeps them from being bored; and there is also an implicit internal audit, if you want, because you know, sooner or later, your office neighbour’s going to be looking at that code.”

What we can learn

AdNovum is a classic example of a dialectic process in action. The original thesis was to allocate people to projects in a top-down way. Arn offered a compelling antithesis: create an internal market, give people a chance to put themselves forward. What we see now is a genuine synthesis: a blend of formal tools and informal discussions that gives a voice to the individual developers while still leaving the management team in control.

Synthesis doesn’t emerge without some creative tension. As Arn observed, the internal-market model works well in a high-octane environment of skilled and competitive people. But it also has its flaws. Software development needs ‘grinders’ as well as stars, and the grinders often get a raw deal when their performance is analysed in objective terms. Also, linking bonuses too closely to ranking lists means encouraging people to focus on a narrow set of behaviours, whereas there are many soft dimensions of performance, from helping colleagues to taking on the messy project no one else wants.

So, little by little, AdNovum softened the edges of its innovative project management model. Project leaders made provisional bids and then worked them through in a meeting, with give and take; ranking lists became more informal, present in the back of everyone’s mind but not on the Intranet for all to see. Moreover, bonuses were based on subjective as well as objective criteria. “As time goes by,” observes Crowden, “you automatically start realising that certain things are of use and certain things are nice to have and certain things are not really that important; and then, eventually, without really noticing it, it’s like evolution.” The internal-market system now works through a conversation; and that’s really the only way it can work if a firm is trying to coordinate the needs of multiple clients and multiple programmers.

What’s next for AdNovum?

There are clearly scalability challenges with a project-planning model of this type, as you cannot have a single weekly planning meeting for a thousand-person organisation. So, in February 2011, the company introduced a new structure with smaller development pools organised by speciality – the project managers and business analysts are in one pool, the application engineers are in another and the security and integration engineers are in a third. Allocations are now done in the same way as before – understanding and catering to individual needs – but focused around these pools. There are also plans to put in place a new HR software system to manage the initial allocations as well as to provide better functionality for individual programmers to update their CVs and development needs online. For his part, Arn is now trying a similar pool-based approach in a much larger organisation – UBS.

“In principle, we are today still doing the same thing as in the beginning,” observes Crowden, “but we adapted the Cluedo meeting to our changing needs. We probably focused more on the tooling for project allocations; and the aspect of the internal marketplace, in terms of individual skills and experience, has become a secondary priority. The changes we are looking at today may allow us to return to our roots with an internal market but on a new level and based on eight years of experience.”

With offices in Silicon Valley and London, the Management Lab works with leading-edge firms to help them create tomorrow’s new practices today. Visit our InfoCentre or website.

Share: Facebook
Facebook Twitter
Twitter LinkedIn
LinkedIn Email
Other Print
Print Newsletter

About Us