How better demand forecasting could have saved Dean Foods a lot of spilt milk

Posted on May 13, 2010 by Gene Tanski

With demand forecasting,

it’s easy to get carried away with the technical, the theoretical, or just plain get lost in the weeds. I often see practitioners who zero in on specific facets of forecasting and demand planning, and in so doing, lose the complete and holistic approach that can deliver measurable business improvements.

In thinking about the big picture, I believe we all need to keep in mind that there are real and significant consequences to getting demand planning and forecasting right or getting them wrong. This story in the WSJ was a critical reminder. Dean Foods announced that first quarter earnings are down 43%. The CEO said that they are facing increased pressure from private label products, which might handicap growth and profitability into the future. The most troubling aspect to me? No real numbers regarding how much of a handicap, nor any clues as to a competitive response – in other words no substantiated forecast detailing the competitive response and supporting the credibility of Dean’s future value.

Dean Foods Co.’s first-quarter earnings fell 43%, and the company’s chief executive warned Monday that some business may not bounce back as retailers use private-label products to lure customers.

Chairman and Chief Executive Gregg Engles said gains by private-label producers are accelerating, noting that in some regions, a half-gallon of Dean Foods milk costs more than a gallon of an unbranded product. Retailers routinely use discounted milk and bread to drive foot traffic, and private-label food products gained share as consumers traded down during the recession.

I’m bringing this up not to kick dirt on Engles and his team, but to highlight some very specific implications of forecasting. A more comprehensive approach to a forecast would have revealed these icebergs much earlier and given them the insight to effectively react.

Better data, process, and technology
I posit that there are three things that could have kept them out of this stew: appropriately use of data (Comprehensive point of sale from retailers, demographics etc.), a better process (having sales people and account managers contributing market intelligence) and an appropriate technical platform. (My advice on the technology is obvious, but it looks like they’re stuck with a big German solution.)

More time to control expectations
Earlier detection would have allowed Dean’s to alert analysts and investors to the trend and the potential hit to their earnings. It also would have prepared them for a competitive response — critical to maintaining credibility with the investor community.

Did I mention better technology?
Appropriate software would have alerted Dean’s to the trend and allowed for comprehensive “what-if” analysis of plausible responses — as well the margin impact of each potential response — so that the company could have quickly coalesced around a competitive strategic response.

Dean’s is struggling, and I would argue a large portion of it is due to lack of a sufficiently robust capability around demand planning and forecasting which as we know is really the tactics behind understanding the market and your customers. The impact is real: their stock lost 25% of its value in the early hours of trading today. More importantly, what is their response? What are the options? Will they be profitable? Will the retail customers respond? Critical questions for the Dean Foods team in the days to come…

Gene Tanski, Demand Foresight featured on Bold Ventures Radio

Posted on May 4, 2010 by Gene Tanski

I’d like to thank Jean Creech of Bold Ventures Radio for including me on her Wed., April 28 broadcast. Jean’s show airs in north Atlanta at 1620 AM and around the country at www.americaswebradio.com.

Bold Ventures Radio focuses on entrepreneurs, and it was my honor to be included as we discussed demand planning, forecasting, technology and business trends — as well as the genesis of Demand Foresight and our supply chain management software. I found this to be a stimulating experience and look forward to the chance to chat with Jean again.

You can listen to the whole episode here. I come on after the first few minutes.

IT’s fiduciary responsibility for business performance: Best-of-breed versus the failed ERP approach

Posted on April 22, 2010 by Gene Tanski

A few months ago, I wrote about IT and the competitive disadvantages inherent in an ERP approach. The short version: one-vendor suites are not a competitive differentiator and they’re not a business strategy. With the market requiring CIOs to be more business- and customer-driven than ever, I’m frankly surprised that salvos like Rick Veague’s (CTO, IFS North America) still get any serious credence. In the back-and-forth about best-of-breed and ERP, I appreciate the multiplicity of perspectives. But I say without equivocation that the assertions in this article are just plain wrong, top to bottom. Unfortunately, this seems to be the default perspective for too many CIOs.

In recent years though, there has been a growth in the number of stand-alone software solutions — “best of breed” applications as they are called — that threaten to roll back the progress experienced by manufacturers these many years. These software products deal with only a segment of the enterprise, inhibiting the free flow of communication and reducing efficiencies. The more these software products proliferate, the more expensive and confusing enterprise technology becomes, and he [sic] more difficult it is to coalesce data for reporting and manage enterprise-wide security.

First of all, the big ERP suites are fundamentally a collection of best-of-breed functionality that was acquired piece by piece from competitors who outperformed them in a particular process or function. ERP vendors are essentially marshaling functions and integrating them in the background in the same way a best-of-breed, or bolt-on, strategy would. You’re just paying a hefty premium for the brand name.

ERP cannot provide the best in all things, all the time. Instead of empowering critical teams to seek the best way to do business, everybody is yoked to lowest-common-denominator performance. My thinking is uncomfortable and inconvenient for many CIOs who know that big suites make their life easier and provide a dependable supply of embroidered shirts and free rounds of golf. But it ultimately comes down to this: once a company is strategically clear about how they’re going to compete, they should then go find the technology to execute, giving their people the absolute best tools they can to beat the competition. ERP vendors routinely blur three critical areas of focus: best of talent, process and platform. This is old-school, reflexive protection of IT empire. It’s locked-in thinking that is killing companies’ ability to compete.

Selected facepalm moments from this article

Late in the article, Veague highlights a number of disadvantages of best-of-breed. On the whole, I found these objections were shallow, ticky-tacky or low-level. What they collectively miss, and it’s a big miss, is the imperative that all technology should directly support your strategic initiatives. One vanilla system cannot possibly help your company do this. If you have an ERP that’s sub-optimizing critical functions that your team needs to better compete, I can guarantee you that it’s costing you a lot more than the growing pains of integrating a best-of-breed ever would. I’m talking about revenue, cost-efficiencies and EBITDA. If you can’t point directly to how IT is driving all those factors, your IT isn’t doing its job.

That’s the standard we hold ourselves to, and the linchpin of our company’s value proposition: we guarantee that our software will deliver outstanding performance (like reducing forecast error by 25%) and a bottom-line impact (minimum 5% increase in EDITDA) that would offset any issue listed below 100 times over. It is this type of performance that underlies the strength in best-of-breed. Companies need to understand it, utilize it, and let their IT help them be more competitive rather than dummied down and made to be like everyone else. But on to Veague’s points…

Integrating best of breed solutions requires the work of systems integrators, adding cost and substantially extending implementation time. These efforts are typically unnecessary with a Suite application.

Ain’t necessarily so. Today, integrating can actually be faster and cost less.

Reporting and access to information can be more expensive using a best of breed solutions because corporate information is spread across multiple applications and platforms.

Technologically, this is completely wrong. This has been proven time and time again. For example, having a data warehouse that supports best in breed/performance architecture bears a cost that is certainly no more, and possibly less expensive than an ERP.

A best of breed solution can increase costs for supporting technology acquisition and maintenance costs.

Only if mismanaged. The amount you pay for installs, maintenance and upgrades of an integrated suite are going to more than erase any savings you get here.

Different best of breed solutions tend to have distinct or unique security models, and that means it is harder to maintain security and privacy across an integrated collection of products.

Sorry, but this sounds like garbage to me. There are many applications that will exactly mirror the security model from your ERP and/or financial systems.

Usability and the ability to collaborate are often diminished with a collection of best of breed solutions because users that must work cross-functionally must learn different user interfaces and systems. This is in contrast with a Suite product that offers a consistent, well-thought-out user experience.

I think that this is fundamentally wrong. If you have a big ERP suite with a forecasting module that is there because, well, it just comes with the rest of the bundle, your people aren’t going to use it to the degree they they should. Your best performers aren’t going to adopt some perfunctory, clunky what-not just because it’s integrated. However,if you give them a tool that is tailored to the way that they work and they see how it will make them better, then they’ll see the value and usability will go up accordingly. If you go down to the lowest common denominator, people don’t get deeply involved and they won’t collaborate anyway.

There are a lot of companies out there that are less competitive than they could be because thinking like Veague’s is taken for granted. I’ll be blunt about it: any officers at any company who swallow this stuff wholesale are abrogating their fiduciary responsibilities to their teams, their bosses and their shareholders.

Competitive Advantage: Best Practices for Delivering a World Class S&OP, Forecasting and Demand Planning Capability

Posted on April 2, 2010 by Gene Tanski

This might be more about me than you wanted to know, but guess what? The new season of Dancing with the Stars is on. My girls and I are enjoying being all snarky about outfits and physical coordination; we are jointly rooting for Evan and Nicole and I, of course, have a personal fondness for Pamela. Watching dancing that’s alternately enjoyable and gut-wrenchingly horrible naturally gets me thinking about — wait for it — forecasting and demand planning. Are most companies good at forecasting? How would they rank as contestants on FWTS (Forecasting with the Stars)? Okay, I won’t take that one any farther.

Just the same, most of the research from the likes of AMR/Gartner indicates — and our experience at Demand Foresight certainly reinforces — the idea that the current state of forecasting and demand planning is in a state similar to a young couple learning how to dance. They might currently be focusing on spins or twirls or lifts, but it isn’t until it all comes together in fluid, integrated simpatico that it really works and makes the dancers and the audience happy.

The perfect steps
What, then, is the perfect incarnation of demand planning and forecasting? Ideally, it would be a way of day-to-day working in which:

• Every professional within an enterprise understands that the forecast is the most important business tool to get right and use correctly.

• It would be supported by simple and straightforward processes and enabled by intelligent and helpful technology.

• It would be measured.

• It would form a large portion of compensation.

• It would be an articulated focus of the organization and its leadership in pursuit of its vision and strategy.

Unfortunately, in most enterprises, the current state of demand planning (modeling and managing the demand side of the business) falls far short of unison, confidence and grace.

For example, we as a business community don’t even know what to call it: Demand Planning? CPFR? S&OP? Forecasting? Sensing? Guessing? We discussed that in an earlier set of blogs – hopefully our definitions make sense.

Timing the steps: what are the time frames for demand planning and forecasting?
Some of the terms above and/or approaches are delineated by time frames (someone mentioned demand sensing the other day) — short-term forecasting focused on minutes, hours and days. S&OP, on the other hand, is purported to focus on mid-term time frames — 6 to 18 months out. Then, of course, there is Strategic Planning (the providence of highly paid prognosticators, researchers and people afraid of getting their hands dirty actually making something), which focuses on issues years out (minimum 18 months and longer) — capacity requirements, brands, market positioning etc.

In my mind, all of these time frames are included in a successful demand planning and forecasting dynamic. They feed into one another and are dependent on each other. Successful forecasting around a new product launch (very near-term forecasting: “Do consumers like this color?”) will impact strategic thinking about what new products in which to invest. Investments in capacity will drive new marketing activities, which will promote sales volume that needs to be forecasted 2 to 4 weeks out, if that capacity is domestically sited. I think you get the point: fundamentally, the process needs to include what is going to happen 10 hours from now, 10 months from now and 10 years from now — all are part of successfully integrated demand planning.

I also want to spend a little time with this oft-occurring question: what is the right view point to drive demand planning and forecasting? Sales forecasts? Marketing forecasts? Customer forecasts? Financial forecasts? These are different pictures of demand based on inputs from professionals with specific responsibilities, external factors, and business requirements (“We need to have this market share or margin,” etc). The answer is not that one of these viewpoints is correct, but that each of them has value in the true picture of demand and need to be included into the process of continuously quantifying demand. The hard question is how to include all of these points of view.

You are well aware of the appropriate technical platform for enabling this work…right? But what is best process? The best process is the one that works, the managers are willing to enforce etc. All the same stuff has been laid out in numerous books and courses through out the years. But I think there are a few other keys…

Ownership of forecast accuracy
One key we have seen is that regardless of the process, one person or group does have to own the forecast and its accuracy. One way we have seen this work well is for each group to have its own forecast (our software happens to support multiple forecasts across an organization): sales will have a forecast, as will marketing, maybe even the customers’ forecast can be integrated, too. Then each is locked according to the process time frame, i.e., first Wednesday of the month).

Now, let’s say the planning group owns the forecast. They will perhaps host a meeting (a consensus meeting?) and get everyone to compromise enough to generate one consensus forecast that can be sent to operations and/or purchasing. Or if planning really owns it (and is measured on it and compensated by it), they will make the decisions unilaterally. The key is: each forecast submitted from each areas of focus can be measured against actuals to get a specific view as to forecasting accuracy. And that will help drive future process improvements and allocation of responsibilities.

In order to keep everyone focused on the process — do not forget, this is really, really key — do not forget to have every one of the input groups, from sales to finance, have part of their compensation predicated on forecast accuracy. There are number of ways to do this, and perhaps we will share some experiences in a future post. But the big point is to have forecast accuracy drive compensation in some form or fashion. And this makes perfect sense if you believe in the importance of forecasting for bottom line performance and competitive advantage over time.

When this holistic version of forecasting is working well, it is like a really enjoyable dance. When one professional from sales makes a change to a short-term forecast, they understand they are also impacting the planning group’s 13-month rolling budget view and the strategy group’s 3-year investment plan. During the monthly forecast cycle — time should be spent 12 or 13 month out — think what this would save (time, resources) for annual planning process.

And like the beautiful, organic dance routine that the new dancers want to master, this is a goal for demand planning and forecasting in companies that are going to thrive in the future. In fact, it is imperative that everyone understands and accepts the responsibility for as accurate a forecast as possible.

This means accounting for as many time frames as possible. It means taking into account the many different perspectives within and without an organization; Utilizing the many forms of data currently available to organizations — both internal and external. It means making demand forecasting a strategic imperative within the organization, tied directly to the successful attainment of the corporate strategy and vision. And that means making sure compensation is tied to the accuracy of the execution level forecast. And by the way, the technical platform to enable this is key and I happen to know a solution that can enable this nicely.

P.S. Go, Pamela!

Where’s the love for bootstrapped companies?

Posted on March 24, 2010 by Gene Tanski

Just finished reading a WSJ article about 50 high-potential, venture-backed firms. Good article and good companies. However, this got me thinking: Why this focus on venture-backed firms? If an entrepreneur has no other options, then funding from a venture source is obviously fantastic. But it comes with a huge price.

Bottom line, the original entrepreneur team that works with a venture group ends up with 10% or less of the company they founded. While their ownership is shaved down with each new tier of financing, their control over the vision and day-to-day management of their company is similarly curtailed. More and more, the management team (vetted and installed by their VC backers) focuses on meeting investor-driven goals that are centered on sheer dollars rather than strategy or vision.

There are obviously many more detailed arguments, both pro and con, concerning venture firms. But it seems to me that a lot of people look at venture backing as a substitute for their own due diligence: “If a venture firm has decided to invest, they must be good company/have a good idea!” And this is the gist of my issue with the article mentioned above — where are the garlands for the American heroes who bootstrap their companies?

Here is the difference between venture-backed companies and bootstrap entrepreneurs: bootstrappers believe so much in what they are doing, that they will risk everything — kids’ college funds, the homes they live in, their retirement funds — to make it happen. Their belief is strong enough that they won’t risk being compromised by short-term financial pressure applied by VCs whose only eye is on the quickest ROI possible. Bootstrappers forgo quick returns and make sure that their customers have provable results that corroborate their own wild convictions. They believe so much in what they’re building, they’re not going to dilute the eventual reward by giving it away to money people.

As you look at the future of demand planning, supply chain management and forecasting, keep your focus wide enough so that you’re seeing more than features, functionality, and dollar signs. Also look for the companies who shun venture capital and make it work because they know they’re creating huge value for their customers — and will do anything to nurture that value.