Human Centric Analytics (HCA) would seem a natural paradigm to use when defining, measuring, and reporting on Environmental and Social Governance (ESG) as it ties together a holistic societal approach in tackling social and environmental sustainability, and the need for reporting and measuring the levers of change in business. What methods can we use to help us to do this? How can we ensure that behaviours are embedded alongside measurement protocols such that change becomes real, and not just a ‘tick box’ exercise, hence creating more sustainable and less risky business throughout the corporate landscape? In this months blog our founder and chair, Christina Phillips, ponders these questions.
As Eli Goldratt famously said “tell
me how you measure me and I’ll tell you how I’ll behave”.
So, why do we need to be so
concerned about analytics when it comes to ESG: To quote David Harris, Global
Head of Sustainable Finance, Data and Analytics at LSEG. “The finance and
investment community can drive the solutions to secure a sustainable and net
zero emission future, to achieve this, there is a need for robust data and
analytics.” This is true of finance, but it is also true within the firm. A
recent report by Accenture on behalf of the UN interviewed over 100 CEO’s to reimagine
global pathways to resilience growth and sustainability. They spoke of how AI, analytics,
and robotics were changing the way procurement and supply chain management can
be done. They also saw a need for artificial intelligence and predictive
analytics to be used for data driven scenario/risk planning.
Deloitte’s ‘Blueprint for a green
workforce’ report found climate change and decarbonisation to be the highest
priority for firms regarding their strategy and operations. They point to a persistent
digital skills gap unable to satisfy a need for embedded analytics throughout
supply chains and development of robust but updateable KPIs. This suggests a
need for analytics development that prompts knowledge growth and skills
acquisition alongside the data and analytics design pathways.
Operational Research, described as
‘the science of better’ by The OR Society, has many tools for acquiring
decision parameters and developing metrics. It is also good for running
scenarios using different techniques and even has many methods for handling
messy problems! It also has a growing set of human centric frameworks that have
codified procedures to bring humans and analytical tools together.
The key to unlocking effective
human data interaction is to ensure a bridging of understanding. After all
there are usually three, more or less aligned, truths to data: what the data
say, what the data are supposed to say, and what people think the data say
(technically multiple truths since everyone could have their own version!).
This data multi-verse needs bridges to understanding, which means opening up multiple
perceptions as well as data sources before attempting to bring these together
often across diverse stakeholders.
Human Centric Analytics (HCA) is a
design paradigm (a way of doing design) that follows the principles of human
centred design. The idea is to grow knowledge as you design and put into use
the analytics required. This needs methods to foster active participation and
discussion by the humans involved and ways for them to interact with the analytics.
It also needs sympathetic and broadly experienced analysts who can call on
multiple methods/dimensions/visualisations to find the ones that make sense in
the stakeholder’s context.
When we worked with stakeholders in
a complex manufacturer to create a simulation for scenario exploration, there
were many occasions when we had to create a space for dialogue between the
data/analytics and the stakeholders (both individually and as a group). People
needed to understand what their experience looked like in terms of data and
analytics to be able to relate to the parameters we would need to make a model
work. Without this understanding they could not have agency in the decisions
around model parameters or understand what the model was saying. This meant I
had to listen to what they were experiencing and relate this to data sources,
for instance I created clustering based on their context and knowledge of the
products being made. Once I had clusters that we all agreed through a process
of statistics/visualisation and discussion, I was able to do further analysis
and visualisation such as histograms and box plots that used colours they had
chosen. Stakeholders recognised their experience in the way the statistical
models were behaving. They were also absorbing a refresher course on statistics,
learning about the available data and its systems, and learning to use visual
tools.
This story brings us to the next
step in HCA that of creating an acceptable level of granularity. This is where
the measurable parameters are decided upon, for instance the time step, the
occurrence rate, the level of detail, or the number of functions to include.
When we took the simulation model up to the SLT it had to become a highly
simplified conceptual model with the key indicators measured and the highlights
of the more complex modelling noted. While that same model was being developed
with staff whose processes it was modelling, we had to iterate a couple of
times before we got the granularity right and then again before we got the
modelling right. The process of achieving an acceptable level of granularity
created knowledge and prompted behaviour change by opening up insights into
manufacturing process and creating networks of involved stakeholders. This
happened for everyone SLT included.
The rest of the HCA process is much
like human centred design: there are likely to be multiple iterations of
development, the needs to be a chance to reflect and discuss, to be creative
and to allow solutions to emerge in use/for use. The process is non-linear and
should not be expected to be and it is best to figure out ways around obstacles
after having first created ideal solutions. Don’t expect everyone to agree but
hope for the best alignment possible!
Roles are important when using HCA.
It is a good idea to have a champion from senior management who believes in the
project and wants to see it work, someone prepared to commit enough time and
energy. One also needs to make sure the evidence is recorded well, so diary
keeping can help as well as either audio/video recording/note taking. Of
course, there are always the trusty stick it notes, rich pictures and maps
(causal, cognitive, feedback, mind map). Collating this evidence well and
ensuring it can be, and is, accessed at a future date is a crucial part of HCA
since we have to take this qualitative human data as seriously as our systems
data.
I believe that HCA is one way to
help us create firms that are able to address the serious challenges involved
in achieving net zero and the worldwide Sustainable Development Goals (SDGs).
As Tariq Fancy (former senior ESG analyst) points out “We’re running out of
time: we can no longer answer inconvenient truths with convenient fantasies”. Our
KPIs and the people who measure them need to ensure that we are not just
ticking boxes but are actually making organisations more responsive and in so
doing making our future lives and jobs more sustainable.
If you would like to know more
about human centric analytics, or if you are interested in trialling an HCA initiative in your sustainability
endeavours please get in touch with Dr Phillips: c.j.phillips@ljmu.ac.uk