Most companies are not operating at a world-class level when it comes to experiments, models, data, decisions, working with uncertainty, and analysis. Analyzing big data, structured data, unstructured data, even not very much data is tricky. Most companies do a fairly poor job of understanding what their data tells them. It's easy to mistake noise for signal. It's easy to miss false negatives and believe false positives. Studies show that about 70% of decisions are made poorly given the information available, resulting in either type 1 or type 2 errors. What if your analysis and decision is high quality, but then your PR people mess up the message and send out a poor press release? It happens more often than not. Are your people trained in the flaw of averages? Are you doing experiments with single variables and randomly assigned control groups? Do you have a decision-support group that helps make big (or small) decisions? Are you aware of the cognitive biases that affect decisionmaking?
We don't think executives should be data scientists. We think they should hire good data scientists and understand how best to work with them. We provide five key support service desks for:
This group is in charge of the model that a) everyone contributes to and b) everyone uses for forecasting and talking about the future. This is a crowdsourced model that will be far more accurate than any one person's view of the future. It's also a quantitative model with APIs that can be plugged into any particular model anyone in the company is working on.
We don't expect your people to be decision scientists. We provide decision-science at all levels, so when anyone in the company needs help with a decision, this team springs into action and helps create a framework for that decision. Our experience shows that people spend far too much time and money gathering the wrong information for making a decision. Very few companies bother to compute the value of information they bring to a decision. This team helps go straight to the heart of the matter and make decisions under uncertainty that fit in with what everyone else in the company is doing. It also goes a long way toward fixing meetings.
Most companies say they are doing experiments, but when we ask them about randomized control groups, they draw a blank. We help design experiments properly in the first place, gather the evidence, then draw conclusions based on what was learned. This is a very tricky area of data science that most data scientists don't have training in.
Early Detection, Early Response
Companies that have a dedicated team to watching the horizon for "black swan" events that could have a huge impact on the organization tend to be much more prepared with scenarios and information for real-time decisionmaking. It's not about planning. It's about being aware and having the resources to jump on a situation the second it arises. In 2014, for example, many companies were surprised by events like Ukraine, the fall of the price of oil, the rise of the dollar, the announcement of QE by the European Central Bank, and many others. This team is constantly on the lookout for surprises and can try to mitigate their effects quickly.
This team helps keep the trainers and the teams fresh, with new content, new tools, and new approaches learned from the cutting edge of the lean and agile movements. There is always something new to learn, and continuous improvement is a big part of business agility. Don't fossilize your learning back to the time when the agile coach came and left - keep it alive, up to date, and constantly improving through outside learnings from the lean and agile communities.
We help you hire, manage, and maintain these services. Or, we can provide them directly, via on-site staff, on-call staff, APIs, and Software. Then we train managers in how to get the most out of these services and not override them with gut feel or expediency. These must be available company-wide, even if your company has 80,000 employees. Contact us to learn how these services will improve your bottom line.