The Successful Business Case
Good business case analysis will not eliminate risk in decision making, but it can reduce the uncertainty to a minimum, measure what remains, and provide the means for minimizing risk as the action goes forward.
Business Case Risk Part III:
Can We Really Trust These Results?
Do you trust your business case results enough to act on them? Should you? It's one thing to project good results from projects, products, or partnerships--assuming things go as planned. It's another thing to trust the predictions.
Part I of this letter discussed pitfalls on the road to getting the accountability and "comfortable certainty" that decision makers are looking for. Part II showed how good risk and sensitivity analysis can measure the likelihood of seeing different outcomes from you actions. In Part III below, we decide whether or not to trust the business case results.
The Devil is in the Assumptions
The key to making that decision lies in understanding the assumptions behind the cost and benefit estimates. Remember that good business projections should not be viewed as the output of a "black box" predicting system or a high-powered analysis that very few people understand. Ideally the case builder wants to position business case results like this:
"I have drawn several pictures of the way the future may work out (scenarios). These are detailed and concrete scenarios based on assumptions about many factors, each of which comes with some uncertainty. However, if the assumptions stand, these results certainly follow." All the uncertainty, in other words, lies in clearly stated assumptions, not in the way that costs and benefits are projected from them.
Whether or not we trust the projections depends on what we know and believe about the assumptions. Below is a simple example, using a single cost item estimate and three assumptions. The approach is the same, however, when analyzing a full cost/benefit cash flow statement from the case (for more on what belongs in a business case, and why, please see one of our books by clicking the images below.
Can You Trust This Estimate?
Management in a network services company is deciding whether or not to launch a new customer service offering. They will go forward with the service only if they are confident it will be profitable. This calls for a decision support business case, with a large number of estimated cost and benefit items.
Here is their approach for estimating one cost item, annual labour costs for service delivery. The estimated is based on just three assumptions.
The average values are the most likely values in the eyes of case builders using the expected averages:
Annual labour cost estimate =
12 months * Average No. of calls per month * Onsite time * Labour rate =
12 * 100 * 4.00 * £80 = £384,000
The road from assumption to cost estimate is out in the open for all to see and evaluate. If the assumed values turn out exactly at the averages given, no doubt about it, the annual cost of labour will be £384,000.
Should management trust that estimate? Remember, these decision makers are highly risk averse. They will go forward only if they have confidence in the projections. If the cost of labour really turns out to be more than £500,000 the service becomes unprofitable. How likely is that?
Run the Future 10,000 Times
Risk analysis addresses such questions by applying Monte Carlo simulation1 to a financial model. In this case, the "financial model" is just the formula above.
Notice first the ranges of possible values for each assumption: the case builder expects an average 100 calls for month, but in fact the number could be more or less. However, the case builder is very certain that it will not be less than 70 or more than 135. The other assumptions also can--and probably will--differ from the expected value, but the case builder is very sure they will fall within the ranges given. With the e assumptions viewed as ranges of possibilities, and with a few more assumptions about the likelihood of different assumption values,2 we can use Monte Carlo simulation to "run the future" 10,000 times and summarize the different cost estimates that appear.
Here is what the case builders found with their first-pass simulation run3 There is a 50% probability of seeing an actual cost of £384,000 or more. That's no surprise: we already knew the expected average.
But these risk-averse decision makers are more than a little uncomfortable with a 20% chance of having actual costs hit £500,000 or more. Moreover, the 90% confidence interval for the cost estimate is very wide: £240,000 through £530,000.
Can they do anything to build confidence in seeing that labour costs stay low enough to make the service profitable? Some very clear guidance comes from another result of the simulation exercise: sensitivity analysis:
While performing the risk analysis, the simulation program was also keeping track of the correlation between each assumption and the forecast result (here, the cost figure). This chart tells us that the dominant assumption, by far, in controlling different cost results, is the assumption about time spent per service call.
If case builders can reduce the uncertainty in the "time per call" assumption, they can reduce the uncertainty in the projected cost. Here, it turns out that more discussions with the service product manager involved, and with the service personnel training manager, led the business case builders to narrow the range of "near certain" average onsite time requirements from 1-7 hours to 3-5 hours (very occasionally there may be a much longer or much shorter onsite call time, but they are very sure now that the average onsite time will be in the narrower range, 3-5 hours).
Case builders also reduced the range of estimates for "average labour rate" from £60-£100/hr to a much narrower range of £75-£85 per hour. Very sure now that the assumptions will fall in the narrower ranges, the case builders re-ran the simulation program and produced this result:
The most likely result is still £384,000, but now the probability of an actual cost at £500,000 or more is virtually 0, if you can believe the new assumptions about the assumptions!
Furthermore, management can be 90% confident that the actual cost figure will fall between £344,000 and £422,000. If they trust these results, they are ready to implement the service.
Should you Trust These Results?
The question of trust all comes down to what you believe about your assumptions. If the (now narrower) ranges of possibilities for the assumptions stand, then the probabilities in the final summary graph can be trusted.
--- --- ---
`Building An Excellent Business!` is available for immediate purchase by visiting the secure Cavendish eStore online at: http://www.cavendish-mr.org.uk - customers can download this e-book in PDF Acrobat format immediately after purchase.
More Economics and Business Inspiration:
`Accelerate with Impact` -
by Colin Thompson ISBN: 978-1-84549-289-2
Accreditation: UK Registered Learning Provider:10025755
Note: About the Author Colin Thompson
Colin is a former successful Managing Director of Transactional/Print Manufacturing Plants, Print Management/Workflow Solutions companies and other organisations, former Group Chairman of the Academy for Chief Executives and Non-Executive Director, helping companies raise their `bottom-line` and `increase cash flow`. Plus, helping individuals to be successful in business and life in general. Author of several publications, research reports, guides, business and educational models on CD-ROM's/Software and over 400 articles published on business and educational subjects worldwide. International Speaker and Visiting University Professor.
The Successful Business Case - Business Case Risk