Case Study of Capgemini Invent: Making Crossdisciplinary Collaboration Work in a Large Corporation
I had a chat this week with Kary Bheemaiah about enabling cross-disciplinary collaboration to innovate within a large tech corporation and the role of team-structure, incentives, and trust in doing so.
Kary works as an Engagement Manager at Fahrenheit 212 (F212), an innovation strategy firm that got acquired by Capgemini in 2017. To reposition their brand from ‘IT systems partner’ to ‘innovation partner’, Capgemini also made other acquisitions and nested them in their innovation arm: Capgemini Invent(CapInvent). Among the 6 departments of CapInvent, F212 leads the innovation & strategy (I&S) department.
Kary is also the incoming CTO of the Future of Tech(FOT) division of Capgemini Invent. A big part of his new job is to link together the departments to address silo issues and help F212 build products & services by that I&S dept. comes up with by stewarding different stakeholders along a shared journey.
I am including here some of the most interesting excerpts from our discussion. You can also listen to the audio recording of the chat. Excerpts included here are not a transcript of the call and may include notes from my discussion with Kary outside this call.
Kary emphasized that the structure of the teams in CapInvent is kept modular to mitigate silo buildup. The flexibility to assemble a cross-disciplinary task force allows CapInvent to build solutions without limiting the scope across which a problem demands to be innovated upon. The modular structure of these task forces is complemented with a connective tissue that enables easy overflow of ideas, feedback, and resources from one ‘cell’ to another. Kary’s job as an interdepartmental liaison in Capgemini Invent is to build this connective tissue. This structure is similar to the ‘Small-World Network Structure’ I highlighted in my contribution(Slides 37-38) to the Yak Collective report —Don’t Waste the Reboot. A partial snapshot is included below.
A cross-cluster link (or connective tissue) acts as a porous gate to filter interesting signals from intracluster noise to share knowledge across clusters to harvest the low-hanging fruits of the economy-of-scope. Kary stated that his liaisoning role involves being a probe and a filter to find the exciting things happening within the company, in the market, and in the wider tech ecosystem to pick and promote the signals that are high-potential.
When a problem becomes too overwhelming for one cluster(discipline), the adjacent ones jump in to share the load. Or when a cluster finds a small hack that just works really well, they share that to adjacent clusters which then might be able to apply that solution (or a slightly repurposed version) to a different problem in a different context.
Leveraging the economy-of-scope this way multiplies the return of innovation (ROInn) happening inside corporate divisions with impermeable boundaries and a rigid hierarchy (that resemble a scale-free network structure shown above). After the walls of cluster-jargon are punctured and connective tissue gets built across clusters, shared vocabulary emerges and combinatorial innovation gets accelerated. The Small-World Network(SWN) structure maintains the coherence benefits (and the resulting efficiency) of clustering and also mitigates the chances of having blind-spots that make the overall structure (and the company) under-prepared and fragile to shocks, by leveraging the economy-of-scope. Kary highlights that granting competency-based official rank and authority are important to align incentives before a large corporation can leverage this structure.
Incentive & KPIs
Senior executives drive a big part of the innovation decision-making process in a large corporation. Their biggest incentive is to ‘not get fired’. How can the outdated and fuzzy KPI metrics (and the appraisal/approval procedures they govern) be bypassed by a cross-disciplinary task force?
Kary laughed while noting that “those wishing the KPIs in a large corp to disappear will keep waiting for that to happen, it’s not going to happen anytime soon”. But he also suggested there is a need for a new KPI that is more relevant and common among the clusters than the old KPIs. For him, that new KPI is Cash. According to Kary:
Using cash as the KPI is useful here. Quantifying signals, showcasing impact and using the shared vocab based on Cash as KPI is a way of getting the sense of urgency you feel into the minds of others. This is a collaborative effort and respect has to be given to all as one can give direction, but the path has to be created by the teams.
Fuzzy KPIs that innovation & strategy teams tend to have and vanity metrics that technical teams like to brag about are boiled down to precipitate into the number that makes sense for everyone: Cash.
Incentivizing cross-disciplinary innovation doesn’t have to be limited to cash though. It can also be more specifically defined. At Capgemini Invent, in the biweekly CxO council meeting, everyone is asked to bring new proposals that serve the goal of showcasing Capgemini as ‘the innovation partner’.
The ‘Deal Flow KPI’ takes into consideration the new projects & services that are launched by an executive along with new partnerships made with startups/firms in the outside tech ecosystem.
“Execs are now told to work on creating department specific (New Service Offerings) NSOs. There are guidelines on what is an NSO (needs to use new tech, solve a problem via efforts from 2 or 3 Invent departments, address sectoral challenges, etc.). Execs and teams are reviewed based on their participation and success. The forthcoming annual report (June 2020) will talk about NSOs and Applied Innovation networks. As it hits shareholders, executives will onboard in reply.”
Aligning efforts need aligned incentives. But even with aligned incentives, the friction encountered during the building of intercluster connective tissue is significant due to vastly different values and Working styles of different disciplines. The friction needs a lubricant to be minimized. That lubricant is trust.
Friction & Trust
It’s not uncommon to hear jokes with connotations that people in one department don’t respect the people in another department that’s on the other side of the business. Sales guys say “Engineers are too happy building features they want, they don’t know what the customer wants.” On the other hand, Engineers usually joke that the “Sales & Marketing guys don’t usually build anything of value, they only talk about what engineers build.”. These jokes hint at an undercurrent of the lack of respect that exists among disciplines and the lack of interdepartmental connective tissue that is lubricated by the trust.
At Capgemini Invent, the solution to this problem was building connective tissue via workshops that explain the innovation & strategy process to the technical team along with the jargon and on the other hand, workshops that educate the innovation team about the technical tools and production constraints at an introductory level.
These workshops engendered among participants more respect for people in the other team and gave them shared vocabulary to use while brainstorming or prototyping a new product/service.
Kary shared an example of a project going downhill because of misaligned expectations and how it was refactored and rescued by him and his team by setting up a common context for the different stakeholders of the project.
The following snippet is how Kary described it in our discussion outside the call:
“We had a project where we were to build a new supplier procurement platform for a client that is one of the largest banks in Europe (and the world).
Client’s problem: The current procurement process is inefficient, is almost analog, does not help to find the right kind of suppliers, and is impacting the cost of operations and the quality of the partnerships.
Innovation Teams response: Let’s build a new procurement platform that is based on NLP.
Why? The client had the budget and very specific demands that made using an existing provider inappropriate. Also owing to the size and volume of the bank, they have an aggregator effect in EU markets.
Approach selected: We will design the platform based on the aggregator effect of the client. Using NLP, we’ll be able to better categorize suppliers, their offerings and how it fits the strategic intentions of the bank.
1. The innovation teams’ knowledge of NLP was amateurish at best. The platform capabilities were theoretical and the analysis it was supposed to do was feasible but required a significant allocation of technical expertise.
2. This led to the second problem. Designing a new platform takes time (surveys had to be done, automated due diligence processes had to be analyzed, etc.) By the time to template of the platform was created, it had been 6 months. this was when the Technical teams became involved. They were shot into a moving train and were not as committed/could not allocate the NLP experts to the build what was presented to the client. It led to a lot of volatile conversations.
Solution: I was jettisoned into this conversation at this time (it was my first month). Having some experience in the subject (I worked as a director of research at a startup where we made a NLP based search engine) I realized what needed to be changed was the deployment plan.
The following communications ensued:
Sit down with the innovation team and explain why what was promised could not be delivered in one shot. NLP and ML capabilities build off each other based on new data coming in. You needed to have a more ‘obliquity’ based approach where you start small, train the algo’s and add features as the efficiency of the algos grows.
Sit down with the technical team lead and broke down the execution plan into smaller bits. What data do we to train the algo’s on, how will new data be collected, cleaned, processed and used to achieve the objectives outlined progressively? Etc….
Sit down with client key stakeholders (and tech and innovation team leads) and explain why this obliquity approach is a more sensible approach – reduces bad matching risk, more accuracy, less reputation damage, etc..
This led to a change in the action plans, milestone setting etc, which the client was comfortable with.
Having a willing client also helped assuage the tension between the innovation & tech teams. More collaborative workshops were organised and both sides got to see how they go about doing their work.
The outcome was the understanding that the tech teams need to be involved in client conversations from the very beginning – the first interview. Not later down the road. We’ve now made it a mandatory process (even though it seems evident).
You are always going to have silos. If you try and break them, it’s a never-ending effort & pointless. Instead, build more connective tissue between them so that they understand each other’s roles but still maintain their specialty/identity. Tribes and hierarchies will always be there. More communication can make the silo membranes more permeable – and could even lead to some trust and respect.”
You can listen to the audio-recording of the chat I had with Kary here on my Substack newsletter.
If you want to reach out to Kary Bheemaiah, LinkedIn is his most preferred channel. Find him here.
If you want to chat with me, drop me an email to email@example.com.
Back to Writings page
My Newsletter: Daybrew's Note
Join my newsletter to stay updated with what I'm working on: