Decisions, Decisions — A Trip Down the Rabbit Hole…
How our modelling fits into the decision-making process
A lot of us are in the decision-making business.
Sometimes these can be making small, daily calls. Other decisions look at options in the bigger picture — large infrastructure developments, government policy interventions or business strategy. Regardless of size or scope, all of our decisions, or lack of decisions, have an impact.
As a product owner for Arup’s City Modelling Lab, I spend a lot of time thinking about how those big decisions get made. Or, more pertinently, I spend time heading down the rabbit hole of how better decisions get made.
We’ve found that the better decisions tend to be the ones that are better informed, consider the options in sufficient depth, the ones that are ultimately made with a greater level of confidence.
Confidence can be a tricky, fragile thing. Unless, that is, you’ve backed it up.
How do we support decision-making?
“Begin at the beginning,” the King said, very gravely, “and go on till you come to the end: then stop.”
Take your average decision — an external or exogenous factor dictates a decision be made or the decision maker reaches a point, through the assessment of options, where they are comfortable enough to be proactive. Put simply, either something forces your hand or you have the confidence (hopefully through due diligence) to be proactive. Either you make a call or it gets made for you.
To avoid the Cheshire Cat effect we’ll only focus on one aspect of this here, proactive decision making but perhaps we’ll look into reactive decision making in the future.
The focus of our work is to support the latter concept, providing decision makers with a level of confidence commensurate to their needs, and importantly providing transparent and clear evidence to support this confidence.
Life has few certainties, decision making even fewer. Every decision is different and this difference is overwhelmingly driven by the adequate consideration of context.
At the lab, we try to use this unique, often complex, context to inform how we design and deliver our products and services to Clients.
For simplicity in this chat we’ll split our Clients into two tranches — Public and Private.
Regardless of whether they have similar or differing goals, they all need to make decisions.
However, the level of analysis, option consideration, confidence and evidence required to make and support those decisions can vary significantly. There’s lots of reasons for this and I’ll look to illuminate some key ones here.
In preparation for this approach, as a team, we must make sure that we’re flexible and agile enough to fit our option simulation and insights to this differing spectrum and scale of reasons — without compromising effectiveness or efficiency.
Let’s follow the white rabbit a bit…
We could consider that we make decisions proportional to our confidence — but in reality, that’s a little simplistic.
In the real world, we can’t proportionally fit our decisions into arbitrary confidence buckets, otherwise we’ll have decisions with not enough confidence — creating disproportionate risk — and (bear with me here) decisions with too much confidence to the point that they are over-engineered.
The equation also includes risk vs reward, so we must explore the decisions we make, the environment in which those decisions are made (including external requirements for evidence and justification) and the value we seek to offset vs the risk in making the decisions.
Curiouser and curiouser…
Let’s say we still have that bag of decisions we need to make, and we’re a Public entity — what levels of confidence are we going to need to make those decisions?
Overall, our observation and understanding has shown us that due to the rigours of public office, decisions typically require a great level of evidence and ‘confidence’.
This is not necessarily linked to making better or worse decisions, that’s another whole conversation, but is as much as any other reason linked to the need for transparency, clarity and discoverability in the process of considering and communicating options.
Acceptance of the option, route or policy decision being made is typically widely distributed, meaning that the spectrum of acceptable evidence to sometimes arbitrarily justify these are also diverse. This manifests itself in a decision distribution that trends towards ‘a lot of decisions requiring a high level of optioneering, confidence and evidence’, and a long gestation period of achieving that confidence in making a decision.
Practical observation notes a similar level of analysis being undertaken on both high-value and medium-value decisions — a point of potential risk and inefficiency that is often criticised — but given the frequency and volume of decisions required by public entities this systemic issue is one not simply remedied.
Accepted systems and processes for making decisions in a public-context take a long time to develop, and even longer to validate and formally accept. The processes are created to make lots of decisions, at a higher level of certainty, with low risk and with high transparency.
On the flip side, private actors and entities must contend with external dynamics such as competition and timing. They must be ready to balance risk and reward with opportunity costs.
To ensure that decisions have a ‘positive impact’ for their business (a wide definition), they must not undermine value by spending too much time, resource and efforts overcoming uncertainty to an excessive degree.
Take too long and you will miss the boat.
In this case, spending too much time and resource considering options to discharge risk before deciding may, at best, risk-reducing net value or upside.
As such, private entity decision-making distributions may be less skewed towards increased confidence requirements, with businesses being willing to take ‘an acceptable level of risk’ to achieve a ‘commensurate or positively disproportionate value return’.
Interesting, but where does simulation fit in here?
There are no certainties in life, but there are a lot of repeated behaviours and patterns.
Harnessing these patterns, modelling these behaviours and considering a wide array of potential scenarios and options we can simulate outcomes that will provide a far higher level of insight to traditional analysis.
We simulate outcomes, to inform decision paths, assess options and provide insight and confidence. In creating our models and simulations, we align with these decision distributions to make sure the outcomes we’re providing are suitable for the decision or action being undertaken.
“Would you tell me, please, which way I ought to go from here?” “That depends a good deal on where you want to get to.”
At the City Modelling Lab, we have a number of approaches that have organically developed over time, on an open-sourced basis in collaboration with our Clients and Partners.
The level of development of these models and simulations is commensurate with the decision making system and the insight sought.
We nominally refer to these developments in Greek-alphabetical terms or sometimes in line with the UK Government Digital Services agile delivery guidelines (we’re pretty loose with it but it works for this discussion…).
The public-client development timeline, and associated value perspective is relatively straightforward
However, there are some important points to note.
As development, verification and validation increase — confidence increases and typically so does value in terms of the ability for our simulations to support public-facing formal decisions making.
Value accrues as certainty of outcome is generated, this means that although there are valuable opportunities early, the highest value add for the Client, is in later stages of development.
Early discovery developments are a fantastic way to augment existing, accepted decision-making support tools and provide fresh perspective for dialogue and engagement around potential options. Mature, validated and widely accepted models and simulation techniques take a long-time to develop and agree upon but ultimately provide significant value to a large number of decisions of varying values.
To address this, the City Modelling Lab develops Alpha models (https://medium.com/arupcitymodelling/innovate-uk-building-an-alpha-abm-as-a-response-to-covid-19-1de5b11d51e6) to investigate early-stage insights, and works collaboratively and extensively with clients to build simulations beyond this, such as our multi-annual programmes in Ireland (https://www.arup.com/projects/transport-infrastructure-ireland) and New Zealand (https://www.arup.com/expertise/services/digital/city-modelling-lab).
These multi-annual approaches will clearly take a longer time to develop, validate and gain acceptance — much of which is not reliant on the technical efficacy of our techniques themselves but on the socialising of our works within the associated industries.
As we progress through development, these models add incrementally further value and we hope that they continue to support our clients’ decision-making frameworks evermore in the future.
Private bodies can potentially derive more value from quicker, earlier developments of simulations as part of the overall approach to decision-making within their businesses.
An open and transparent approach to development can support robust and clear discussions around risk, value and confidence provided by option simulation activities leading to potentially business-differentiating insights over competitors who are not undertaking such analysis.
Armed with this intelligence, and a clear understanding of the value potential and confidence in the outcome businesses can make timely decisions, consider potential scenarios for varied impacts and modelling anticipated constraints to implementation.
Some decisions will still require a higher level of confidence, and we find our more emergent simulation techniques can support accepted decision making support tools in this regard, but where there is a less skewed, more symmetrical distribution of decision-confidence we believe there is significant value to be gained through our approaches to modelling and simulation.
Considering this in the development timeline, it is our perspective that the development of Alpha level, and beyond Alpha simulations are of significant value to private entities.
In contrast, the extensive validation, socialising and acceptance of further modelling and simulations may be too resource and time-consuming; ultimately undermining the net value of any such activities.
That is not to say that such approaches cannot add significant value, but this would have to be in response to either a few number of incredibly high-value, high-confidence requirement decisions (such as financial regulatory volatility simulations) or a high-frequency of business critical functions (such as private aeronautics or space operators for example).
On the whole, we believe that private businesses such as logistics providers, ecommerce, utility providers and private transport network operators could benefit from the insights provided by early-level simulation, and we are positioning ourselves accordingly.
She generally gave herself very good advice, (though she very seldom followed it).
In light of these considerations, the Arup City Modelling Lab is constantly looking to match its development to the needs of its clients and we must deliberate a key factor of our advice accordingly — reliance upon our simulation and associated technical advisory services.
Combining our simulations with our specialist domain knowledge of our Clients’ sectors, be that transport, water, energy or planning we add a unique level of additional value and rigour to Arup’s advice — providing an overall service that differentiates us from our competition.
“No, no! The adventures first, explanations take such a dreadful time.”
