Scenario forecasting

A fundamentally different approach to judgmental forecasting is scenario-based forecasting. The aim of this approach is to generate forecasts based on plausible scenarios. In contrast to the two previous approaches (Delphi and forecasting by analogy) where the resulting forecast is intended to be a li ely outcome! here each scenario-based forecast may have a low probability of occurrence. The scenarios are generated by considering all possible factors or drivers! their relative impacts! the interactions between them! and the targets to be forecasted. "uilding forecasts based on scenarios allows for a wide range of possible forecasts to be generated and some e#tremes to be identified. $or e#ample is it usual for %best&! %middle& and %worst& case scenarios to be presented! although many other scenarios will be generated. Thin ing about and documenting these contrasting e#tremes can lead to early contingency planning. 'ith scenario forecasting! decision ma ers often participate directly in the generation of scenarios. 'hile this may lead to some bias! it can ease the communication of the scenario-based forecasts! and lead to better understanding of the results.

The Delphi method

The Delphi method was invented by (laf )elmer and *orman Dal ey of the +and ,orporation in the -./0s for the purpose of addressing a specific military problem. The method relies on the ey assumption that forecasts from a group are generally more accurate than those from individuals. The aim of the Delphi method is to construct consensus forecasts from a group of e#perts in a structured iterative manner. A facilitator is appointed in order to implement and manage the process. The Delphi method generally involves the following stages1
1. A panel of experts is assembled. 2. Forecasting tasks/challenges are set and distributed to the experts. 3. Experts return initial forecasts and justifications. These are compiled and summarised in order to pro ide feedback. !. Feedback is pro ided to the experts "ho no" re ie" their forecasts in light of the feedback. This step ma# be iterated until a satisfactor# le el of consensus is reached. $. Final forecasts are constructed b# aggregating the experts% forecasts.

2ach stage of the Delphi method comes with its own challenges. In what follows we provide some suggestions and discussion about each one of these.-

Experts and anonymity

The first challenge of the facilitator is to identify a group of e#perts that can contribute to the forecasting tas . The usual suggestion is somewhere between / and 30 e#perts with diverse e#pertise. 2#perts submit forecasts and also provide detailed 4ualitative justifications for these. A ey feature of the Delphi method is that the participating e#perts remain anonymous at all times. This means that the e#perts cannot be influenced by political and social pressures in their forecasts. $urthermore! all e#perts are given an e4ual say and all are made accountable for their forecasts. This avoids the situation where a group meeting is held and some members do not contribute! while others dominate. It also prevents members e#erting influence based on seniority or personality. There are suggestions that even something as simple as seating arrangements in a group setting can influence the group dynamics. $urthermore! there is ample evidence that a group meeting setting promotes enthusiasm and influences individual judgement leading to optimism and overconfidence.3 A byproduct of anonymity is that the e#perts do not need to meet as a group in a physical location. An important advantage of this is that it increases the li elihood of gathering e#perts with diverse s ills and e#pertise from varying locations. $urthermore! it ma es the process costeffective by eliminating the e#pense and inconvenience of travel! and it ma es it fle#ible as the e#perts only have to meet a common deadline for submitting forecasts rather than having to set a common meeting time.

Setting the forecasting task in a Delphi

In a Delphi setting! it may be useful to conduct a preliminary round of information gathering from the e#perts before setting the forecasting tas s. Alternatively! as e#perts submit their initial forecasts and justifications! valuable information not shared between all e#perts can be identified by the facilitator when compiling the feedbac .

Feedback
$eedbac to the e#perts should include summary statistics of forecasts and outlines of 4ualitative justifications. *umerical data summaries and graphical representations can be used to summarise e#perts5 forecasts. As the feedbac is controlled by the facilitator! there may be scope to direct attention and information from the e#perts to areas where it is most re4uired. $or e#ample the facilitator may direct the e#perts5 attention to responses that fall outside the inter4uartile range! and the 4ualitative justification for such forecasts.

Iteration
The process of the e#perts submitting forecasts! receiving feedbac ! and re-viewing their forecasts in light of the feedbac ! is repeated until a satisfactory level of consensus between the e#perts is reached. 6atisfactory consensus does not mean complete convergence in the forecast value7 it means that the variability of the responses has decreased to a satisfactory level. 8sually

two or three rounds are sufficient. 2#perts are more li ely to drop out as the number of iterations increases! so too many rounds should be avoided.

Final forecasts
The final forecasts are usually constructed by giving e4ual weight to all e#perts5 forecasts. The facilitator should eep in mind the possibility of e#treme values which can distort the final forecast.

Limitations and variations

Applying the Delphi method can be time consuming. In a group meeting! final forecasts can possibly be reached in hours or even minutes 9 something almost impossible to do in a Delphi setting. If it is ta ing a long time to reach a consensus in a Delphi setting! the panel may lose interest and cohesiveness. In a group setting! personal interactions can lead to 4uic er and better clarifications of 4ualitative justifications. A variation to the Delphi method often applied is the :estimate-tal -estimate method:! where e#perts can interact between iterations. The forecast submissions can still remain anonymous. A disadvantage of this variation is the possibility of the loudest person e#erting undue influence.

The facilitator
The role of the facilitator is of utmost importance. The facilitator is largely responsible for the design and administration of the Delphi process. The facilitator is also responsible for accommodating feedbac to the e#perts and generating the final forecasts. In this role the facilitator needs to be e#perienced enough to recognise areas which may need more attention than others! and to direct the attention of the e#perts to these. Also! as there is no face-to-face interaction between the e#perts! the facilitator is responsible for disseminating important information. The efficiency and effectiveness of the facilitator can dramatically improve the probability of a successful Delphi method in a judgmental forecasting setting.

Forecasting by analogy
A useful judgmental approach often implemented in practice is forecasting by analogy. A common everyday e#ample is the pricing of a house through an appraisal process. An appraiser estimates the mar et value of a house by comparing it to similar properties that have sold in the area. The degree of similarity depends on the attributes considered. 'ith house appraisals! attributes such as land si;e! dwelling si;e! number of bedrooms! number of bathrooms! and garage space are usually considered. 2ven thin ing and discussing analogous products or situations can generate useful (and sometimes crucial) information. 'e illustrate this point with the following e#ample.-

Example 3. Designing a high school c!rric!l!m

A small group of academics and teachers were assigned the tas of developing a curriculum for teaching judgement and decision ma ing under uncertainty for high schools in Israel. 2ach group member was as ed to forecast how long it would ta e for the curriculum to be completed. +esponses ranged between -< and =0 months. (ne of the group members who was an e#pert in curriculum design was as ed to consider analogous curricula developments around the world. )e concluded that >0? of analogous groups he considered never completed the tas . The rest too between @ to -0 years. The Israel project was completed in < years. (bviously forecasting by analogy comes with challenges. 'e should aspire to base forecasts on multiple analogies rather than a single analogy! which may create biases. )owever these may be challenging to identify. 6imilarly we should aspire to consider multiple attributes. Identifying or even comparing these may not always be straight forward. As always we suggest performing these comparisons and the forecasting process using a systematic approach. Developing a detailed scoring mechanism to ran attributes and record the process of ran ing will always be useful.

" str!ct!red analogy

Alternatively a structured approach comprising a panel of e#perts can be implemented as proposed by Areen and Armstrong3. The concept is similar to that of a Delphi7 however! the forecasting tas is completed by considering analogies. $irst of all a facilitator is appointed. Then the structured approach involves the following steps.
1. A panel of experts "ho are likel# to ha e experience "ith analogous situations is assembled. 2. Tasks/challenges are set and distributed to the experts. 3. Experts identif# and describe as man# analogies as the# can. !. Experts list similarities and differences of each analog# to the target situation& then rate the similarit# of each analogies to the target situation on a scale. $. Forecasts are deri ed b# facilitator b# using a set rule. This can be a "eighted a erage "here the "eights can be guided b# the ranking scores of each analog# b# the experts.

6imilarly to the Delphi approach! anonymity of the e#perts could be an advantage in not suppressing creativity but could hinder collaboration. Areen and Armstrong found no gain in collaboration between the e#perts in their results. A ey finding was that e#perts with multiple analogies (more than two)! and who had direct e#perience with the analogies! generated the most accurate forecasts.

F!t!res thinking
In academic circles futures thin ing is associated with futures studies. $utures studies is described as an interdisciplinary %collection of methods! theories! and findings& (Biller! 300=! p.@) that helps people to Cthin constructively about the future5 ("ell! -..D cited in ,odd et al

3003! p./). It has also been summed up as %the rigorous art of imagining&! with applied e#pressions across a range of fields from big business to education for sustainability. The emergence of futures studies The emergence of futures studies is generally credited to the late -.D0s and early -.@0s. Three groundbrea ing boo s of that era include Alvin Toffler5s -.@0 Future Shock, Alain Touraine5s -.@- The Post-Industrial Society. Tomorrows Social History1 Classes, Conflicts and Culture in the Programmed Society! and Daniel "ell5s -.@= The coming of Post-industrial Society a !enture in social forecasting. 6hell5s -.D.--.@0 )ori;on Eear Flanning project led the way within business by constructing possible scenarios for business -/ years into the future. 'hen one of its least e#pected scenarios (an unprecedented dramatic oil price hi e) became reality less than three years after the project began! 6hell5s readiness to act swiftly encouraged other companies to ta e up scenario development. )owever! after a decade of popularity! enthusiasm in scenario planning is said to have waned in the -.<0s. 6ome commentators blamed its demise on misunderstandings about both the process and purpose of future scenarios as %the method became used over-simplistically! with confusion between forecasts and scenarios& (van 'ieringen! 6ellin and 6chmidt! 3003). Forecasting versus scenarios This point about the difference between forecasts and scenarios is an important one because it hints at a range of Cways of thin ing5 associated with scenario development specifically! and within futures studies in general. 6cenarios have been described as %internally consistent and coherent descriptions of hypothetical futures& often at least 30 years ahead of current time ((2,D! no date). $orecasts tend to be more concerned with accurate prediction. )owever! various typologies of both forecasts and scenario methods e#ist! with some overlap! and these are just two of the analytic tools used to imagine possible futures. Different assumptions about the future )ow and why specific tools are designed can be partly e#plained by the inds of assumptions that drive different futures thin ing projects. 'hile some futures thin ers might assume that the future can be e#trapolated from past and present trends! others e#pect G or at least create space for G radical and unforeseen transformation. Hi ewise! some might prioritise articulating a "refera#le or "ro#a#le future for people to then plan towards! where as others open up "ossi#le futures in a far more e#ploratory and divergent manner. These lines of difference raise deeper 4uestions about how! and for what purpose! %the future& is conceived. 6ome suggest that the goal is to ma e preferable futures more probable! by visualising what we want to create and then committing to it. Another common goal is to be to be prepared for the remotely possible (as in the case of 6hell). (thers might argue that distinctions between possible! preferable! and probable futures are counterproductive! because how we act in relation to them in the present defines their li eliness in the actual future.

+iel Biller (300=)! who wor s on the (2,D5s 6chooling for Tomorrow project! provides an educational perspective. $or him the aim of futures thin ing is %neither prediction nor advocacy& but %to pursue the un nown&! an idea closely associated with comple# systems thin ing. Biller suggests that when people are supported to become creative and rigorous futures imaginers! they come to realise that the future is not something that will happen to them tomorrow but is being created #y everyone today. )e refers to futures thin ing as a %navigational tool& for %changing the nature of decision-ma ing in the present& in ways that are %embracing IofJ comple#ity! interdependency! and real-time decision-ma ing&. In other words futures thin ing helps people to thin about and design for emergent outcomes within comple# systems9outcomes that! by definition! cannot be fully nown. Futures thinking for shifting thinking (pening up a space for people to envisage possible futures can be a %motivator for getting unstuc & ((gilvy! 300D) by enabling new insights into interrelated patterns that could be reoriented to disrupt past trajectories and create alternatives.