Duh-mand?

Demand is like electricity:  it's powerful, quick moving, and essential to any business.  But demand is also like electricity in that everyone "knows" what it is, but very few people really understand how it works.  Just try to explain the difference between 1 volt and 2 volts to somebody who isn't an engineer.  If you succeed at that, try to explain the difference between a volt and an electron volt.

Unlike electricity, demand is really easy to explain in the abstract (it's intuitive: people want the stuff you're selling), but hard to quantify in a practical sense.  How much demand is there in a market (separate this from, "how much did we sell")?  How much demand will there be a year from now?  If this were easy, everyone could predict demand and Wall Street wouldn't be any fun at all (like it's any fun now?)
 

Demand

Quantifying aggregate demand is important very early in a company's life because the business needs to have a convincing story about the size of the market opportunity to attract VC investment.   Ironically, at this stage there may not be a market yet (no money is changing hands because the product doesn't exist) so there won't be solid data.  If you're lucky, industry analysts are putting out SWAG numbers.  Even if your new company is in an existing market category (providing a clear replacement / substitute product), it's hard to know how demand is going to change when the new product is introduced.  Ten years ago, who would have believed that 20 million Americans would be willing to pay $3000 and up for a TV (in '98, it was almost impossible to spend more than $1500 on a TV, and the average was probably $400). 

Coming up with these visionary forecasts is hard, requires significant domain knowledge, and is subject to a lot of error.  One of the toughest issues in the formative stages of a market is, "what's the boundary of this market?"  This is a very squirrelly issue with products that are components or subassemblies, or products that are substitutes for services.  Sometimes what you think of as a discrete product is perceived by the customer to be a feature of a much larger product or service.  If you don't have a clear or realistic definition of the market space and its relevant segmentation, demand estimates won't mean much no matter how big they are.  Further, the estimate you come up with will be the total available (or accessible) market (TAM) for all vendors...but how much of this total will your company be able to get?  If your product is first rate, if your timing is good, if the product price represents great value, if your sales team kills...if, if, if.  These kinds of projections are more an anticipation of how big demand could get:  a model, not really a forecast.

The other time when quantifying aggregate demand becomes important is after a company goes public.  Wall Street wants to know how much further you can grow and how your market share has changed.  In these cases (for example, automobiles or mobile telephones), there are solid sources of industry demand data, analysts that crunch the numbers, and clearly defendable market shares.  But this only occurs in big markets -- there have to be a lot of
transactions and interested observers. 

It's rare to have solid demand data in product categories below $500 M in sales.  Industry analysts won't put out numbers for your specific product area, and will only quantify large domains that aggregate dozens of related products.  Nobody really knows how big your specific market is, its growth rate, or your market share.  All you really know is how much you sold, and how big your pipeline seems to be.  Beyond the horizon of your pipeline (typically, the length of one sales cycle), you have no solid basis for a demand forecast. 

For many companies between $1M and $1B, investment decisions must be guided by models that extrapolate.  This inherent unreliability is compounded by entrepreneurial enthusiasm, political/budgetary pressures, and math/logic errors.  Many companies in high tech base decisions on forecasts that are undefendable.  Almost invariably, the projections are too high and won't be achieved.

These kind of important questions that come up all the time:

• what's the expected growth curve for our product?
• how do we know Sales isn't sand-bagging the forecast?
• have we maximized revenues?
• is our market share growing? 
• is our competitive position degenerating, or has the market gone soft?
• did deals fall out of the quarter, or were they never there in the first place?
• how much more could we sell if we invested more in marketing or sales?

If you can reliably answer these questions with accuracy for mid-market companies, quit your job immediately and head off to Wall Street, where you will be rewarded handsomely for your skills.
 

Forecasting:  Triple Vision

Most of the time, companies don't put a lot of focus on demand in the abstract.  Day to day, businesses are concerned with predicting how much they can sell this quarter, not aggregate market demand. 

Public companies need to run three sets of forecasts: 

• Revenue -- generated weekly by Finance and the factory, used by investor relations and the board for managing investor expectations
• Bookings -- generated weekly by Sales, used by Sales Management and the President to run  day-to-day field operations
• Product -- generated annually, updated quarterly by Marketing, used to justify investments and to help the factory make supply-chain capacity planning and purchasing decisions.

As all these forecasts tend to jump around for their own reasons (notice they're done by different departments who may not like/trust each other).  It's not practical to synchronize or precisely reconcile them all in real time -- it's rare to fully reconcile all of them more than once a year).  While it's conceivably possible to integrate these forecasts, I've never seen more than two of them effectively converged.  (If you know of someone who's done a completely integrated forecast, please contact me!)

The bookings forecast is the one most people want to know about:  it's the one Sales uses day by day.  In the forecast, the terms "pipeline" or "funnel" are used to characterize the evolution of demand.  But in the complex sale (direct sales rep, $50K+ price per unit), 90% or more of the people who enter the funnel / pipeline will drop out.  So the funnel is a sieve.  Or, more optimistically, the pipeline is a refinery. 

The ultimate problem with bookings forecasts is their inherent unreliability and inaccuracy:  they are usually based on untested models, use extrapolation extensively, and work with source data that is biased, unreliable, and largely unfiltered.  Many companies find that their forecasts don't get more reliable over time -- they'll be good for a year and then they'll blow up.  Particularly in markets with the "hockey stick effect," it's amazing that forecasts come as close as they do. 

While there are real solutions to forecasting accuracy, they involve sustained effort and annoying changes in several organizations.
 

Finally, a Business Application of Heisenberg's Uncertainty Principle

Unless you're dealing with large numbers of customers and commodity markets, there's no way to fundamentally know how many purchases are going to happen, at what price, and when. 

Let's return for a moment to our electricity metaphor.  It's very easy to accurately predict the behavior of electrons flowing through a copper wire.  But as an ounce of wire contains about 600,000,000,000,000,000,000,000 copper atoms, these forecasts benefit from the law of large numbers.  If you're trying to understand the behavior of an individual electron, Heisenberg posited that you could quantify its position or its velocity, but never both.  Further, he theorized that to actually measure any one electron, your measurement efforts would inevitably change the electron's behavior.

Coming back to business, the parallel is you can forecast the general flow of demand, but it's impossible to fundamentally know what's going on in any particular deal.  Even the customer doesn't know exactly which product they're going to buy, in what quantity, at what price, and at what time.  They only know their overall budget, timeline, and vendor preferences.

Behavioral targeting is the art of analyzing customer behavior to predict the state and evolution demand for an individual customer.  By evaluating customer inquiry patterns (prospect registered for a white paper, then looked at a Flash demo, then attended a webinar, then downloaded a trial copy of the software...), BT engines try to stimulate customers into action when they are most ready to respond.

Great idea, but recent findings have shown fairly similar results with almost nonsensical calls to action (e.g., targeting an auto shopper while they're browsing a cooking site).  The problem here is that the Behavioral Targeting science is based on noise:  the humans you're trying to target are trying to evade the sales cycle, playing hide-and-seek so the pushy sales reps won't bother them.  In other words, Heisenberg.
 

Creating Demand:  A Rant

Returning one last time to the electron metaphor, it is not possible to create electrons.  You can generate a flow of electrons, you can harness them, you can transform them so their behavior meets your needs.  But creating them is God's work. 

Same thing with demand.  Excluding fads, impulse buys, snake-oil, and vice, it is virtually impossible to profitably create demand.  People either need your product / service, or they don't:  it's a matter of making people aware they have a need, making them aware that there are solutions and you offer one, and focusing their desire on your offering.  

So, ask your marketing department to increase awareness, generate leads, and improve conversion ratios.  And ask your Sales guys to work closely with inside sales and marketing to rapidly transform the prospect interest into active sales cycles.  This is the way to avoid getting zapped because you didn't make the quarter.

  Digg This!  

Leads 2.0 -- coming in February

Contents copyright 2008 by DOTnet Consulting, Inc., all rights reserved.
All trademarks and graphics are the property of their respective owners.
Feel free to forward or syndicate this report, but you must include this copyright notice.
Somebody Brilliant™ is a trademark of DOTnet Consulting, referring to its principal.
This newsletter was mailed to you because our records show you asked to receive it.
Click here to opt-out of future newsletters, although that will make us feel very sad.