How predictive analytics can boost
product development
Complex
product-development projects are plagued by schedule slips and cost overruns.
The up-front application of advanced and predictive analytics helps companies
build plans they can stick to.
R&D projects are inherently unpredictable. When embarking on efforts to
design complex things, companies often have little idea how long a project will
take, what it will cost, or what they’ll finally be able to deliver to the end
customer. Their initial project plans are sometimes no more than educated
guesswork.
For example, in an
analysis of more than 1,800 completed software projects, we found that only 30
percent of them met their original delivery deadline and one in five of these
did so by removing or deferring feature content. The average overrun is around
25 percent of the originally planned schedule. The performance of a sample of
over 1,600 integrated-circuit-design projects was even more telling. Over 80
percent of those projects were late, and the average overrun was nearly 30
percent. Moreover, those projects were almost as likely to suffer an 80 percent
overrun as they were to finish on time.
Delays, and the extra
resources needed to counter them, mean higher costs too. The average budget
overrun experienced by a group of factory-automation-software projects we
studied was more than 10 percent. A fifth of those projects cost over 50
percent more than originally expected. Then there are the indirect costs.
Delayed launches mean lost sales, opportunities for competitors to get ahead,
and potentially damaged reputations.
The underlying causes of overruns
We’ve spent more than a
decade investigating the root causes of R&D scheduling and budget
challenges. In that time, we’ve interviewed hundreds of project stakeholders,
including executive managers, technical leaders, and program and project
managers. They highlight many issues that boil down to two primary root causes.
The first root cause is
underestimating the complexity of the project. Managers and engineering teams
are often surprised by the combined impact of all the features and performance
targets and the cost of integration into a finished product. Engineering
intuition tends to be linear, while the cumulative effect of increasing
performance, features, and quality is highly nonlinear.
The second root cause
is overestimating the productivity of the development team. Planners tend to
assume that the issues that befell their previous project would be cured and
that no new issues would crop up. They assume that specifications will not
change and that resources will be available when needed. In practice, of
course, such problems do affect almost every project.
How predictive analytics can help
Until recently, even
companies that understood and sought to address these issues didn’t have
effective tools for doing so. Conventional complexity metrics, like counting
lines of code, story points, or function points (FPs) in software development,
are difficult to estimate before the start of a project, especially one that
requires many sprints from many teams to complete. Story points, by their
nature, are qualitative and team specific, making estimation difficult when
multiple teams are working on the same release. By their very nature, FPs focus
only on function and not the actual effort drivers associated with
implementation and validation, thereby leading to inaccuracies of greater than
60 percent in more than 50 percent of projects that use FP-based estimates. And traditional methods often fail
to account for other external factors, like the programming and development
styles adopted by the development team, multisite development, and the impact
of challenges the team is facing for the first time.
Today, some companies
are adopting a new approach, one that uses powerful data analysis and modeling
techniques to bring new clarity to the estimation of project-resource
requirements. At its heart, the new approach relies on the fact that, while
every development project is unique, the underlying complexity drivers across
projects are similar and can be quantified. If companies understand the
complexity involved in a new project, they can estimate the effort and
resources required to complete it
Doing that is harder
than it sounds. Companies must collect a significant amount of data to
determine what factors really impact project effort. But for practical reasons,
the only useful factors are ones easily measured, consistently gathered, and
known early enough to drive budget and planning decisions. Developing a set of
models, then, relies on an array of advanced analytics, machine learning, and
artificial-intelligence techniques to predict the complexity and required
development effort and schedule in a reliable way. In software engineering, for
example, those models would need to understand the complexity of the system
requirements, the architectures, the testing, and the potential required
interactions with hardware.
Because these
complexity models are based on real data, they don’t make unrealistic
assumptions about productivity. And their estimates automatically incorporate
the effects of the everyday delays and disruptions that development teams must
face. In other words, they take into account not only the complexity of the
project (both the functional and implementation aspects) but also the
complexity of the team environment.
These models can even
identify the productivity impact of changes to working methods. Larger
development teams are less productive than small ones, for example, as they
must expend more effort on internal coordination and communication. The
introduction of new teams, new platforms, or new development approaches can
also hit productivity in the short term, even if they are intended to boost it
over the long haul. With enough industry data, however, the models can see how
these sorts of changes affected productivity in the past and provide a good
estimate of likely future effects.
Better plans, smarter decisions
Armed with such models
and a baseline of productivity levels for similar projects, a company can enter
the current specification and develop higher-integrity plans for new products.
It can then assess the risk of the current plan or create a more realistic
staffing plan along with a good budget estimate and an achievable schedule.
In addition, analytical
models provide a powerful new way to deal with constraints. A company can model
the resource requirements of multiple projects scheduled to run concurrently,
for example, to see if there are any points where those projects will demand
more staff than it has available for a specific role. With warning of such
resource bottlenecks, it can take appropriate action—adjusting the schedules to
separate the peaks in demand, bringing in contractors, or outsourcing part of
the work. Similarly, the models will show if an aggressive budget or timeline
can be made achievable by adding more resources. And if it can’t, the company
can run what-if analyses to evaluate the impact of dropping certain features or
simplifying performance requirements. That allows a much more thoughtful,
fact-based discussion, far preferable to missed deadlines or being forced to
drop features at the last minute because they weren’t finished in time for
launch.
Predictive analytics at work
Predictive analytics
have already have transformed the outcomes of some high-value projects. As an
example, at one company, a project to create a derivative of a newly released
product was originally expected to take just 300 person-weeks of effort. The
project’s planners arrived at this estimate on the basis that 90 percent of the
new design would be carried over from its predecessor. When they reevaluated
the plan using analytic models, they found that the project would actually take
three or four times as much effort. The difference arose because while the
amount of truly new work was small, it was widely distributed and affected
nearly every part of the architecture. That meant significant extra testing and
integration work, which the analytical models identified. Once the company
understood the work involved, it changed its plans, keeping the team that
developed the original product together to work on the derivative, and
ultimately delivering it on time.
In another example, a
company had a tight deadline to complete a new release for a big customer, with
competitors vying for the work. The predictive analytics models showed that
with the company’s current resources and project plan, it was going to miss its
delivery schedule by 50 weeks. That delay would have caused it to miss the
market window and lose a $350 million opportunity. Spurred into action by the
finding, the company took steps to reduce the complexity of its design and
prioritize the scope of the effort, resulting in a project that met the
customer’s minimum requirements and could be delivered on time.
Organizations that
apply analytics and predictive tools to their product-development and
project-planning processes see a dramatic reduction in schedule slippage. And
because they can put the right number of the right people on their projects at
the right time, they also enjoy R&D-productivity improvements of 20 to 40
percent. For companies, that means lower costs and lower risks—a powerful
combination of benefits to have in a highly competitive environment.
By Arjun Balaji, Raghavan Janardhanan,
Shannon Johnston, and Noshir Kaka
https://www.mckinsey.com/industries/high-tech/our-insights/how-predictive-analytics-can-boost-product-development?cid=other-eml-alt-mip-mck-oth-1808&hlkid=b3b1dc5433f0488a82201d030d32dcbd&hctky=1627601&hdpid=a8c82ce0-0d00-4215-b935-54181c0e8b54
No comments:
Post a Comment