demand sensing

Demand Cacophony

by Lora Cecere on July 12, 2011 · 0 comments

The promise was harmony. The delivery is dissonance.

It was to be a unified signal spanning customer’s customer to supplier’s supplier to join, align and guide the supply chain response. For most, it has failed.

Let me start with some perspective. I have followed the evolution of supply chain practices for thirty years. I am an old gal.  Yes, it has been over 30 years.  <My birthday this week is a stark reminder of just HOW long I have been following this evolution.>  As a business practitioner, I built demand planning systems on spreadsheets to help me plan inventory.  As a technologist, I implemented demand management systems for a major supply chain planning company.  I rode the hype of supply chain planning technology cycle with vigor and projected high hopes for supply chain planing.   As an analyst, I have followed the evolution of demand planning processes. I have written about what they are, where they are at, and what they could be. This blog post is my point of view.  I feel that we have let demand planning systems evolve. As technologies have changed, I do not believe that we have rethought the processes to take advantage of new demand signals, better sensing in global markets, or actively shaping demand based on what customers want to buy.  We have not challenged the vendors.  I feel that there is a need to re-architect these processes to seize new opportunities and to deliver on the original promise.

The synonyms for cacophony are jarring, grating, uproar, and clamor.  I find that these words aptly describe the environment of most of my demand planning clients.  Here are the disconnects that I see driving the dissonance.

  • What is the goal?  What is demand? Companies believe they purchased and implemented processes to project demand.  They did.  However, the goal line has changed.  Most have not accepted this reality.  Their internal definition of demand has migrated, and their systems and processes have not.  What does this mean?  The systems that were implemented in the go-go days of supply chain planning, forecast what manufacturing should make using order history.  It is not what the market wants to buy using market signals.  It is far from it.  The devil is in the details. The two goals are not equal.  The demand signals in most organizations are “supply-driven” not “market-driven”, and most companies do not realize the difference.
  • Where do we go from here? How can demand planning systems grow up? The number of potential signals has proliferated. They have less latency.  They better reflect true demand.  In my work, no one argues that use of social signals, sentiment analysis, distributor data, or point of sale data can improve the timeliness and accuracy of the forecast.  The issue is how to use it.  In traditional architectures, there is no where to put it. The models were not designed to use it.  Instead of being able to embrace these new, and varied signals, they are forming a cacophony surrounding traditional demand forecast processes.  People ask the question, “how do I use these new forms of demand data?”  However, when they hear the answer –rethink your demand planning architectures– they are unwilling to take the next step. This redefinition is still early.
  • Confusion.  Are we shaping or shifting demand? The cacophony is propelled by corporate reward system.  Let me explain.  Sales is incented on growth.  Supply chain teams are rewarded for costs. Demand processes struggle to get ownership across marketing and sales because many teams do not want the discipline or the transparency of a market-driven approach.  As a result, many companies shift demand versus shape demand.  This is major.  What is the difference?   Shifting demand is the use of demand shaping levers (e.g. changing price, promotional tactics or new product launch strategies) to move demand from one period to another. Why? There is an organizational struggle.  While it may help a sales executive get a bonus, it increases waste in the value chain and is a barrier for building strong collaborative relationships.  On the other hand, shaping demand–using the demand shaping levers of price, promotion, new product launch, and sales incentives–to increase baseline lift, grow market share and build new markets requires discipline, teamwork, and market knowledge.  Most sales and marketing organizations are not ready to be market-driven.  They are focused on inside-out metrics, not outside-in processes.  Typically the focus is on meeting internal metrics versus maximizing market potential.

<Sidebar:  Before I continue, I want to be sure that we are clear on terms.  I believe that supply chains need to be market-driven.  In my research, I see that the strongest supply chains are connected by horizontal processes from sell-side to buy-side and bi-directionally define go-to-market strategies.  I was a visionary in the writing of demand-driven supply chains, but I believe that the demand driven definition is not sufficient.  Why?  The definition of demand driven is to sense, shape and drive a demand response with near real-time latency of data.  While I believe that sensing, shaping and architecting the demand response is critical, I do not think that is sufficient.  It is missing the bi-directional trade-offs between buy and sell-side market strategies to balance risk and opportunity.  Supply chains are becoming more constrained by supply-side decisions.  Goods at retail are forecasted to increase in price by 20% this year due to transportation constraints, demand has outstripped supply in agribusinesses for the past four years, and raw material prices are a boardroom issue.  As a result, the connection of demand signals horizontally to drive market-driven processes is growing in importance.  There are no market-driven supply chain solutions at this point in time in the market.>

 

What to do?  Five Steps to Take:

 

  • Acceptance. Remember the serenity prayer?  Recognize that conventional demand planning systems are nearing end-of-life. Accept it.  Respect them for what they are, and maximize their use.  Don’t try to use them for something that they were not designed to do.  Accept that you will be living in a demand cacophony for many years.  Learn how to tune into different channels and manually connect the dots. Or alternatively, redeploy demand planning to be a demand-side application and translate what you are going to sell into what you are going to make.
  • Get good at sensing. New systems for demand management will evolve overtime.  However, to get ready to use them, get better at using demand signals.  Get good at demand sensing.  Evaluate new technologies like Terra Technology demand sensing, Signal Demand for demand orchestration, Enterra Solutions for natural language processing, sentiment analysis from SAS, aggregated review data from Bazzaarvoice, and build a demand signal repository from a cadre of vendors (Relational Solutions, Retail Solutions, Teradata, and Vision Chain.)   However, in project planning, realize that this new data does not fit into conventional demand planning processes.  We are not ready for direct integration.  The systems lack the data model, scale and scope to accommodate these signals.  Today, you can only use them to simulate test and learn environments.
  • Be a realist. Recognize today that there is no planning architecture that allows you to use social data, sentiment analysis, downstream data, distributor data, and sales account team input well.  Follow the evolution of technologies in this area and invest in early pilots knowing that the answer is not going to be quick.
  • Build your core. Build strong horizontal processes–revenue management, sales and operations planning, supplier development,  innovation and new product launch processes–to support supply chain strategies.  Build demand management discipline into these processes.  (e.g. In your revenue management processes did you shift or shape demand?  What was the bias and error from management input in the S&OP process?  How effectively did we forecast new product launch?)
  • Develop a market-driven strategy. Prepare. Ask yourself three questions.  “What is the role of the demand signal in connecting sales-side strategies to buy-side strategies to mitigate risk and maximize opportunity?  Why is this important for my business?”  And, what is our roadmap to get there?”
    Would love to know your thoughts.  Please share.
    This week, I am off to SAP Industry Analyst Bootcamp.  Look for my tweets and insights from the conference.  I am trying to understand how columnar store through the SAP Sybase acquisition can help us create better supply chains.  I look forward to getting your comments.

Good Forecasting Matters

by Lora Cecere on April 5, 2010 · 1 comment

The shift happened.  Haiti crumbled.  Chile did far better.  The difference was substantial. The contrast offers lessons on how to shelter cities from future devastating earthquakes.

In a similar vein, the Great Recession shook the core of value chains.  Some supply chains crumbled and some weathered the storm. What made the difference? It is this story that I want to tell. 

The Great Recession

In my lifetime, there has never been one quite like it.  October 2008 shook the foundations of business in a way that I had not seen before. It is the longest recession of my lifetime. 

As credit tightened, supplier relationships crumbled.  Inventory ballooned and factories ceased production. The forecaster was in the hot seat to answer the question of “when will it be over?”  All in all, a very scary time .

Not all were Equal to the Challenge

During this period, I interviewed 28 Fortune 500 companies to understand how long it took them to sense the true shift in demand and how long it took to align their supply chain for the recessionary shocks.  The differences were striking.

Who did it well?

Cisco did it the best.  Burned in the dot.com downturn with inventory write-offs of 2.5 billion, Cisco had taken action to redesign the supply chain to better sense and translate demand.  They sensed the downturn in the first month and had aligned the extended supply chain within 6 weeks.  Their management team understood the importance and acted.  However, for most it was a VERY different story.

 

Leaders sensed the downturn 5X faster and had more quickly aligned the supply chain response.  The interesting fact for me was the capability to sense and translate demand had less to do with technology adoption –the type of system and how it was installed–than attitude and capabilities.  Leaders had the right stuff.  They did three things differently:

Forecast was True North:  Over 80% of companies have implemented supply chain forecasting systems; but, very few have a signal that is true north. True north is a forward-looking compass set on true market demand. Companies with inherent forecast bias were challenged; and companies dependent on sales forecasts were crippled.  The companies that did the best were outward-focused using true market indicators and modeling what-if scenarios based on market drivers.  Forecast models based on orders and shipments were simply not equal to the task.

Outside-in Focus:   Companies that modeled ship-to locations based on selling units out performed.  For most this was too HIGH of a bar.  Why? Over 80% of forecasting systems are built with tight integration to ERP using ship-from locations (distribution warehouses) as opposed to ship-to (channel locations); and modeling is based on the manufacturing unit, which can be VERY diferent than the selling unit.  Likewise, companies that out-performed had replaced rules-based consumption in Distribution Requirements Planning (DRP) with daily statistical modeling using channel demand to calculate required inventories. 

Forecast Discipline:  The companies that did it well were good at forecasting. They had process discipline:  few managerial overrides and when overrides occurred, they held people accountable for bias and accuracy.  These companies had experienced forecasters behind the wheel (as opposed to the new kid just hired from a MBA program), their forecasters knew their products and the market, and they excelled in what if analysis and the ability to use modeling programs. 

Forecasting Matters

  Yes my friends, it is not the system, it is how you use it.  It should be a wake-up call for all. Forecasting matters.  It will matter even more as we come out of the recession.  Case studies are happening right now that will soon be on the front page of the Wall Street Journal.

What do you think makes a difference?  Any stories to share?