Case study: The Fall of Corporations

due to Supply Chain Failure

Introduction
Over the years, Inaccurate Demand Forecasting has been one of the major reasons for supply
chain failure in major corporations.
A precise and efficiently laid out growth demand forecasting system is the key for keeping
businesses afloat in times of massive fluctuations in market demand.
The demand forecasting model can determine whether there is a shortage, a surplus or a smooth
running supply chain engine.

History
Apple was often the leader with respect to market share, in the deeply fragmented PC market of
mid-1990s. That position took a permanent hit in the last half of 1995 due to Inaccurate
Demand Forecasts. Just before the Christmas season in 1995, Apple was to introduce its new
line of Power Mac PCs. Two years before, however, the company had been burned by excess
inventories and production capacity during a similar launch for its Power Book laptops. So this
time, it played things very conservatively which turned out to be the expensive option. When
demand for Power Macs exploded, Apple was caught short for the critical Christmas season.
Forecasts were too low, there wasnt enough flex in the supply chain, and some parts suppliers
developed additional delivery issues. At one point, Apple had $1 billion dollars in unfilled orders
in its system. Unable to capitalize on the market opportunity it had been handed, the stock price
was soon cut in half, the CEO was shown the door, shareholder lawsuits came pouring in, and
Apples market position in PCs took a permanent hit such that it took the IPOD years later to
lead a recovery in the company.
Cisco rode the technology wave of the 1990s to incredible growth, profits, stock valuation, and
prominence for itself as a global business giant. When the tech bubble burst, however, Cisco was
slow to see the slowing demand, and had inventory system and visibility issues that left it caught
unprepared, when its market tanked. In May 2001, as a result it had way more routers, switches
and other gear than it needed; the company announced it was taking a $2.2 billion inventory
write down, probably the largest in history, the company was widely criticized in the business
press which led to stock price being cut in half.

Future Roadmap
To ensure accurate Demand Forecasts, it is imperative to:

Exclude unusual activity that is not likely to occur in the future as even a single unusual,
out-of-scope sale embedded in your historical data can throw off the demand forecasting
accuracy for years.

Lookout for trends that can change demand as it helps avoid any wasteful sales
commitments.

Latest technological advancements like Cognitive computing & Predictive analysis are the
answer to these age old issues.
In supply chain management, Cognitive Computing and its ability to analyze extraordinary
volumes of data, learn from it and make accurate predictions with respect to current and future
market/demand trends; can improve operations, reduce waste, enhance promotions, and
ultimately deliver products more efficiently to manufacturers, retailers, and consumers.
Cognitive computing systems behave like our brain, learning through experiences, finding
correlations, and remembering and learning from the outcomes. They can process
natural language interactions, they have reasoning capacities, and they learn from data that's fed
into it, and from past experiences. What this allows is greater amount of automation and
Predictive Data Analytics Capabilities.
Predictive analytics takes the sea of operation data within an organizations existing system and
turns it into actionable business intelligence which can be used to mitigate risk, optimize future
decisions and create a smarter supply chain. This technology provides organizations with a
clearer view into their supply chain operations and market activity so they can foresee
challenges and respond to them proactively instead of reactively, increasing both efficiency and
profitability. Historical reports alone cant help solve supply chain challenges. Predictive
analytics, however, gives the advanced perspective that allows an organization to understand the
past, present, and future, this type of data insight, can improve forecasting accuracy, optimize
supplier performance, and reduce costs.
An example of such cognitive computing system is WIPROS HOLMES (Heuristics and
Ontology-based Learning Machines and Experiential Systems). Wipro has internally deployed
the system in its help desk. Earlier, when employees logged in technical troubles, they would
have to categorize the problem based on what they thought the underlying problem was. But
30% of those categorizations would be wrong, leading to delays in resolving the problem. Now,
Wipro has fed in half a million of historical technical-trouble logs into Holmes, and the system
itself now categorizes the problem based on its learning from the historical data. The employee
no longer has to categorize manually, and the accuracy has improved dramatically. In retail, the
system would be able to predict what promotion to run at a particular time (it could scan vast
amounts of external data also to come to judgments).

Cognitive computing systems such as these can efficiently categorize relevant historical sales
data, factor in current demand/market trends and make more accurate demand forecasts
preventing supply chain failure.

Compiled by: Anuj Kumar

Source:
http://timesofindia.indiatimes.com/tech/tech-news/Wipros-Holmes-will-challenge-partnerIBMs-Watson/articleshow/45614495.cms
http://www.dnb.com/lc/supply-management-education/predictive-analysis-for-supplychain.html#.VM7xC9KUcg1
http://www.enterrasolutions.com/big-data/analytics
http://www2.isye.gatech.edu/~jjb/wh/tidbits/top-sc-disasters.pdf
http://www.softwarethinktank.com/articles/%EF%BB%BF-top-5-reasons-supply-chains-fail/