MODELING LIKELY UPTAKE OF FUTURE TECHNOLOGIES

1
MODELING LIKELY UPTAKE OF FUTURE TECHNOLOGIES

• Timothy Devinney
• Director, Centre for Corporate Change, Australian
  Graduate School of Management
• Jordan Louviere
• Faculty of Business, University of Technology
  Sydney
• Tim Coltman
• University of Wollongong

2
Who would have guessed ..

• That this technology

• Would become this technology!

3
Or ..

• That this technology

• Would become this technology!

4
Or, that these two ideas would become one .


5
Technology assessment prediction The dilemma

• Rewards to technology developers do not
  necessarily accrue to the best technologies but
  to companies that are best able to match
  appropriate technologies to the latent user
  needs.
• In most cases there are major gaps between what
  firms marketing research (defined broadly) can
  tell them, and what they consider to be
  appropriate scientific developments
• There are limits to marketing research as
  normally construed,
• Many developers are skeptical about the ability
  of marketing research to penetrate into unknown
  futures.
• As technologies become more radical and gaps
  between initial development and market uptake
  lengthen, likelihood of failure rises for two
  primary reasons
• The more radical a technology, the more difficult
  it is to predict how users will react, and
• The longer the time between initial development
  and market uptake, the more likely management
  will be involved more intensively in later stages
  of projects, when projects are least likely to be
  cancelled.

6
The opportunity

• Firms that can forecast uptake more accurately
  should be able to improve innovation hit rates,
  and even slight improvements may position them to
  redefine industries on their own terms, and save
  Ms in sunk RD costs due to

• More accurate design of technologies
  products,
• Less errors in forecasting market uptake,

• Ability to structure timing of new generations of
  technology,

• Being able to influence uptake curve evolutions,
  and

• Knowing when to get on and off uptake curves.

7
Limitations of marketing research methods

• Traditional market research techniques are
  ill-suited to deal with environments in which
• Technology users lack proper contexts with which
  to understand technology applications and their
  potential value to users.
• e.g., 20 years ago how many computer users
  understood or could foresee that the main use for
  home PCs would be internet communication?
• The evolution of a technology matters as much as
  the mature technology itself.
• e.g., current use of communication technology is
  related to past technologies that were available
  to us.
• No technology stands in isolation consumers
  typically use a specific technology in
  combination with complementary technologies.
• e.g., the future of current 3G phones will be due
  less to phones per se and more to 3rd-party
  software that can/will drive applications.

8
A new way forward Information acceleration

• Information acceleration (IA) was developed by a
  team of MIT marketing academics in the early
  1990s who recognised that traditional methods
  failed to forecast uptake accurately due to
• Not providing accurate information to potential
  users about relevant aspects of new
  technologies/products
• Not simulating learning processes associated with
  new innovations and their evolutionary paths and
• Not recognising that individuals organisations
  make decisions choices about technolgies/product
  s, and that market outcomes depend on these
  elemental behaviours.

9
Conceptual background to IA

• How real markets evolve

Pre-Launch
Pre-Awareness
Awareness
M A T U R I T Y ?
Interest
Post-Launch
Capability
Option Evaluation ?? Choice Set Formation
Choose Now
Delay Choice
Never Choose
A, B, , N
10
What traditional methods do

• Traditional Marketing Research Concept Test

Pre-aware Consumers
Best 12 Guesses About Future
How Likely Are You To Do X?
11
Whats missing/needed?

• Need to recognise that the future is a
  combinatorial problem.
• Not 1 future many possible futures many
  possible technologies/products.
• 12 guesses about futures is a sample of size
  12.
• Need to understand model the impact of
  information learning technological evolutions
  on user choices.
• Information has different sources sources have
  different credibilities impacts consumers
  choose/use sources of information differently.
• Need to understand model (combinations of)
  possible new technologies products, not just
  12 guesses of what users might/should want.
• Need to understand how potential users value
  different combinations of possible new
  technologies what they are willing to pay for
  them.
• Need to identify, understand forecast how
  different types of potential users are likely to
  react to different future technologies/products.

12
How to provide whats missing/needed

• The future is a combinatorial problem
• Futures are defined by variables like
  technologies that may/may not exist, complements
  that may/may not be in place, economic
  conditions, etc.
• Information sources of information about
  technologies products are combinatorial
  problems various sources of information can be
  used, information can have different levels
  (including source credibility).
• Possible new products are a combinatorial
  problem
• All technologies/products consist of components
  or features
• Features of products take on various values or
  levels
• Each combination of levels is a different
  product.
• Combinatorial problems have experimental design
  solutions and IA allows us to make use of this
  fact to develop effective forecasting

13
A generic approach to IA structure
Entry Experiment
Context Experiment
Information Experiment
Choice Experiment

• Entry experiment
• How does the consumer go from pre-aware to
  aware
• Word of mouth?
• Solicitation?
• Context experiment
• Is there contextual variation in the market
• Number of competitors?
• Stability of dominant design?

14
A generic approach to IA structure
Entry Experiment
Context Experiment
Information Experiment
Choice Experiment

• Information experiment
• Full information at time t
• Attributes?
• Advertising?
• Testimonials?
• Supporting information?
• Demonstrations/usage?

• Choice experiment
• Attribute trade-offs

15
First example MOBI

• The first example of IA is a voice recognition
  WIFA PDA called MOBI
• MOBI (also known as INCA) was developed by Claude
  Sammit and his team at Computer Science at UNSW
• Limited functionality that includes the ability
  to match seamlessly with ones computer and the
  internet
• Commands include the ability to get news,
  weather, currency information plus make
  appointments, notes and other typical PDA
  functions
• First test was a simple one of the information
  and choice experiments only
• Alternative product configurations where a
  standard PDA with a declining price profile

Information Experiment
Choice Experiment
16
Information conditions
17
Advertising conditions (lifestyle productivity)
18
Media articles (positive, negative neutral)
19
Testimonial conditions (young/old x male/female)
20
Seeing it being used (use seen varied by
condition)
21
Choice experiment structure
22
Estimates of base MOBI choices (market shares?)
23
Product Features PDA and MOBI Impact on
probability of choice
24
Price and cross-price effects
Base demand for MOBI is higher than for standard
PDA but decline based on price increase is
similar
Cross-price effects are similar
25
Role of information conditions (conditional) on
choice
26
Second example bank website

• The second example used more advanced multimedia
  both early in the IA experiment and in the choice
  experiment and utilised the full 4 stage approach

Entry Experiment
Context Experiment
Information Experiment
Choice Experiment

• People received information about the
  availability of the new product in three ways
• The context allowed for security and trust in the
  site to vary
• Information varied advertising, testimonial and
  media attention
• There was a full scale demo of the new product
• Choice was based on switching from existing
  product to variants of the new product

27
Discussion conclusions

• Information acceleration allows for more robust
  and expanded evaluation of technologies, both
  existing, and more importantly imagined and
  radical
• A lens into the mind of the future consumer
• The logical link between choice modeling and
  information acceleration allows for specific and
  accurate forecasting
• Good science, intuitively applied
• The information accelerator allows for quite
  extensive testing and manipulation of the
  environmental context and technology before full
  development
• Future scenarios can be tested
• The information accelerator allows for the
  development of advances in our understanding of
  how to model user needs and in how we
  mathematically/statistically estimate this
• Move beyond simple surveys