#1 Use a single source for portfolio data, both input and output
For many organizations, it’s common that portfolio project data reside in several different systems. Typically, milestones and cost data are stored in spreadsheets, portfolio compilations in a portfolio management system and the output analytics are stored in PPT documents. It is a time-consuming effort to keep track of everything.
To make better use of your data, and without actually doing anything to your data, it should be stored in one single system. This is easier and quicker than to amend slides manually. When using only one source for portfolio data, the most current version is easy to find, and portfolio managers do not have to spend time trying to figure out how input data and analysis are linked – thus there is a consistency from input to analysis. It’s also easier to track the now and then of a portfolio, since the system can automatically present the deltas between versions, both on the input and output.
#2 Include operational risks in the portfolio forecasting
Drug development is risky business. Over the life span of a drug project, there are literally thousands of things that can go wrong that will cause a delay or a cost increase to the project;
Recruitment can be slower than expected
Getting a response from FDA or KOL’s can take longer
Over-recruitment will cause unexpected costs.
Operational risks can be a small nuisance but can also be game-changing events that will call for completely new product strategies. For many organizations, delays are managed in a risk log where project management lists all major delay risks, their impact on project timelines, and most importantly, when the product will launch. Too seldom the operational risk makes its way into the forecasting for individual projects or for the portfolio. Adding operational risk to your forecast thus enhances your overall understanding of the portfolio, and there is a possibility that several forecasts can be created. While there is a chance that several forecasts can be a source for confusion, this can be mitigated with clear communication of the purpose for each forecast. Overall, including operational risk in forecasting increases an organization’s chances of coping with the risk and set up a pre-emptive strategy to mitigate it, instead of finding themselves in the middle of an operational problem they hadn’t predicted with no strategy at all.
#3 Use dependencies between assets as a portfolio overlay
In many organizations, each project in the portfolio is modeled as an independent entity, where in principle, all projects can be launched. In reality there are a large number of dependencies between projects. Here are a few examples:
Project B can only progress if project A is terminated
Project C is a combination product, both project A and project B need to be successful for project C to start
If project A is launched, then project B will be terminated (same indication)
If project A launches before project B, it will have a negative impact on the commercial price for project B
By including project or portfolio dependencies, an organization will be able to create more accurate models that better represent what is happening in the portfolio. Seeing portfolio metrics with and without dependencies can also help portfolio teams and project teams identify new risks that would have otherwise gone unseen. Including dependencies thus helps the organization make better future strategic decisions.
#4 Integrate commercial models with an holistic R&D model
Today, it´s common that organizations have the commercial group separated from the R&D function. The collaboration is done by having commercial group representatives in each development team. The commercial rep will provide the team and the portfolio group with the sales forecast, which in most cases is nothing more than a time-series with a sales number for each year after launch. The forecast is assuming launch at a fixed particular date, a specific order of entry onto the market, and also that the requirements in the Target Product Profile have been met.
Should any of the parameters change, the commercial group will then have to re-create the sales forecast from scratch and then bring that into the team. Even small changes such as a minor delay will require the sales team to recreate the sales forecast. This is a time-consuming process that usually take days up to several weeks.
However, using a holistic organizational model, it bridges the gap between R&D and commercial groups, which gives a holistic end-to-end model of the project. The sales model using parameters from the R&D part of the project model (launch timing, efficacy & order of entry) will be doing so dynamically, meaning that when there are changes to the R&D model, the sales model changes automatically.
By using the same system and even the same model collaboration between the development team, the portfolio team and the commercial group will increase and also will mean that all data is stored in one place. Keeping track of versions will be much simpler and less time-consuming. There will be fewer errors when generating output for slides. One huge benefit is that the turnaround time from making a change to when the new forecast is available will change from days to minutes.
#5 Introduce uncertainty in the portfolio analyses using ranges
It’s too common in organizations that analyses consists of expected values and fixed dates. Each project has a specific planned launch date, a specific cost curve and a specific revenue projection with one value per year in the market phase. Everyone knows there is a huge element of uncertainty in these numbers, however those uncertainties are usually hidden or are represented by unspecified sales scenarios such as a Basecase, an Upside-1, an Upside-2 etc.
By stepwise introducing ranges into the output, an organization can get a better understanding of the risk involved with a project or portfolio. Using ranges, an organization has the ability to identify major risks and can run through these scenarios, prepare for them and mitigate them. By introducing ranges, the analysis becomes a better representation of the reality, as reality is not fixed.
#6 Let the project models reflect the project situation, don’t do just what the template says
In organizations all over the world, portfolio models are built from a template. This is a simplification compared to the real portfolio that may lead to an incorrect assessment of the value for individual assets, and as a consequence, the portfolio forecast may be incorrect. In some cases there are several templates to choose from, but basically they have data on the project phase level, and the sales forecast is one or several time series.
To better grasp the opportunities and the risks in any project, organizations must create its own model that best represents the situation of that specific project. If there is an opportunity for fast-tracking, that opportunity is covered in the model. If there is a risk that an additional study is needed, that is also included in the model. Any project model can be rolled-up into the portfolio and can be used in the portfolio forecasting. To better cope with uncertainties in our decision making, organizations cannot use standardized models, since standardization does not account for project specific risks whereas strategic decisions will be used on faulty indications. Each model has to be specific for any project to better understand its opportunities and risks.