Businesses in any vertical markets are based on tangibles. R&D, investment, cashflow, production workflows, company management, staff skills and efficiency are all identifiable commercial factors needing to be balanced for the business to be run efficiently. Of course, they can be subject to any number of intangible or unpredictable factors. This means any good business leader or manager has to indulge in games of “What if?” to predict what might happen if operating conditions change. Or equipment needs upgrading. Or a global pandemic strikes.
But how do those in the decision-making process know that simulation modelling is the right fit for their business? Debra Slater, Managing Director at Three Media, and an expert in the field of simulation modelling and optimisation, addresses the most commonly asked questions.
- Why does a media business need to consider a modelling tool to manage the “What if” scenarios?
Debra: For most methodologies coming to a logical conclusion as to what may happen will at best list various possible scenarios, document the assumptions and then try and model them in Excel. At worst it is an “educated finger in the air”. Generally, the predictions are variable, sometimes they may get close to an optimal solution, although it isn’t known if it is the most optimal, but more often than not they are wide of the mark.
A much more efficient and accurate alternative is to utilise the latest in simulation and optimisation modelling technology. Software like our product, XEN:Pipeline, driven by artificial intelligence (AI) and machine learning (ML), is able to analyse specific circumstances based on available data, to project likely outcomes and then use these predictions to provide an optimised set of parameters that will provide a guaranteed outcome.
- Who benefits most from running a business simulation model, and how will it streamline a business?
Debra: This kind of analytical power benefits a business in general but specifically makes life easier and more manageable for the senior executives and CxO’s who have a need not only to optimise their organisation but who are always looking to drive new initiatives and implement continuous improvement. The ability to model first and to keep adjusting and optimising prior to making significant changes or justifying investment will allow the executives to instigate an action with confidence and avoid unnecessary expenditure or turmoil.
This technology is also beneficial to the managers running the day to day business so they can assess and predict situations relating to their work and activities. Scenarios can be modelled and impact assessed. The optimisation will generate the right solution to ensure there are no bottlenecks and there is minimal under-utilisation.
- Can you provide some examples where this function can be applied to meet medium and long term objectives?
Debra: There is significant benefit in modelling and optimising controlled changes such as assessing the impact of bringing in new technologies or changing workflows; what investment is needed when taking on and building a relationship with a new client, changing the roster patterns for one or more teams across an operation, financial modelling (including long-term revenue projection, payback on investment and cost breakdown analysis) and validating technical designs which can provide a competitive edge when responding to RFP’s. All of these can be addressed easily and comprehensively.
An area that is probably most pertinent right now is to model and optimise team, department or company wide re-organisations and transformations to remain competitive and maintain sustainability.
- Can it be applied to assist with some of the operational challenges faced on a daily basis?
Debra: On a day to day level, sometimes hour by hour, there are many areas where this can be applied. The results will be optimised to guarantee continued compliance to the many service-level agreements (SLA’s) that exist across a content supply chain and prove there is no impact to the operation. Where this cannot be achieved it will find the least disruptive solution. Examples where this will be beneficial are when some critical kit is faulty and needs to be taken out of service, key resources are sick or a customer wants to process more content with minimal notice.
It also solves some of the conundrums that we have all faced around managing resourcing, not just human but also technical, and that is to ensure that “slack” time is kept to a minimum.
- How can simulated and optimised business modelling reduce the cost base and improve forecasting?
Debra: The core principles of this function is to do just that with the added capability for a user to define what their aspirational cost base should be and what revenue levels they want to meet. The optimisation will then provide the best solution within these constraints. It is not going to be accurate to the penny but it provides informed direction significantly improving forecasting across a team, department or organisation.
- Can modelling guide towards new business opportunities?
Debra: It is easy to see when bottlenecks occur as the operation is noticeably impacted but it is not as easy to articulate accurately the spare capacity, either human or technical. One of the key features of the optimisation is to show where there is this spare capacity in order for this to be exploited. Opportunities can be bid for competitively with the knowledge that profits are still sustainable.
Often new business opportunities are not rolled out as it is felt the impact to do so is not fully understood and the risk is too high. As these new initiatives and opportunities can now be modelled and optimised prior to them being implemented, this will provide the confidence for companies to drive forwards in the knowledge that the most optimal solution has been found, costs are known, and that revenue streams are well understood. The risk associated with implementation is now significantly reduced.
- How do you build and run a model and optimise the results?
Debra: Building and optimising a model within XEN:Pipeline is straightforward through its intuitive user interface (UI). A model is constructed incorporating configuration parameters, which define each step in the workflow or process, and one or multiple data sets which can be legacy or near real time and extracted and manipulated in XEN:.
Once a model is built it can be run for any given period of time in order to evaluate all possible outcomes. A run will output points of interest (POI’s), generated through AI/ML analysis, which will show where problems exist now and / or where they could occur in the future. These predictions are fed into the AI/ML optimisation functions, although users are also able to refine what they would like to optimise and apply any constraints. In addition, a full set of graphs are available across a model to provide additional insight into the overall efficiencies / inefficiencies.
Can a model be run many times with different parameters to come up with the most optimal solution?
Debra: Yes, and that is how it would work in practise. The output of an optimisation can be applied back to a model, re-run over the same time period and the results compared. A model can be run and optimised multiple times, each with different parameters or constraints, in order for you to determine the best solution for the problem.
- Can you run the optimisation in real time, so that a platform is always optimal and SLA compliant?
Debra: This can all be achieved in near real time, where the model is run continually, with pre-defined intervals, to ensure SLA compliance. This is particularly pertinent for operations where workflows and processes are complex, there are multiple SLA’s and it is near impossible to manually predict if they will all be met. It is possible to go deep into the metrics of a record, asset or associated item as it moves through the workflow and discover why, for example, an SLA was missed. Millions of bits of data are generated and stored for each run and these provide the necessary audit trail.
In conclusion, many claims have been made for the implementation of AI/ML in recent years but the ability to run business simulations and discover new efficiencies or create new revenue streams is among the most exciting and useful. Issues like excess or spare capacity and workflow bottlenecks can be identified and highlighted and the true cost of driving efficiencies and implementing new initiatives determined. Now more than ever, with the impact of COVID 19 only just starting to be understood, organisations need to drive change, but they also cannot afford to get it wrong. Now would be a good time to consider a different approach and consider advanced modelling, simulation and optimisation to provide that guarantee.