Smart Data: Keep it Simple

When you set an objective of becoming data-based smart, keep it simple by picking the right horses to ride --these four categories of hard information related to the operation of the company and its competitors. Dont' lose focus, concentrate your efforts; keep it simple.You may want more later, but start here:

The first, actual cost information by product, must be accurate. The term actual is used here to emphasize that the data must be real; many if not most cost systems are not. Before relying on the information generated by your system, check it out in detail. Strip away all the history of overhead allocations; who knows the agenda of the  person making those decisions. If you take your cost accounting accuracy for granted, your decisions might turn out to be as effective as sprinkling fairy dust on your quarterly reports. Armed with accurate cost information you can attack clear targets, like low margin sku’s and slow moving sku’s; almost always many of these hang around in the product line and suck up working capital and administrative costs.

Second is a full tracking of the revenue creation process. If your company sells pre-designed products, track the flow of information from forecasting to the receipt of an order to the succeeding flow of that order information to various departments; the numbering or coding systems used along the way; and the amount of entry or reentry into different manual or automated subsystems such as inventory, invoicing, manufacturing release and shipping.  If your company designs custom products, capture data relating to that engineering process, then to design wins, then to results. Tailor this flow-tracking to your specific business.  Do it manually at first; then streamline the processes and automate; in that order. In every one of my company experiences, some amazingly important and useful revelations have developed from doing this.

Third, compare your operating ratios to your competitors’. Where do you find shortfalls? Don’t assume your competitors are stupid. Don’t assume they’re brilliant. Until you look at the data, who knows? By comparing your costs and prices to theirs, you’ll have a more precise calculation of just how competitive your company is. You can stay the course or set new goals to meet their performance levels. Either way, you’re working with real numbers and zero illusions.

Fourth, every company makes capital and expense investments, most do some analysis before such investment, few carefully track the actual results. When you carefully track the results of investments, your organization will get smarter. When you don’t, you’ll get lost in myths and unsubstantiable stories.

Here's a short story from AMIS that displays a breakthrough from tracking, from beginning to end, one aspect of the revenue creation process, in the process breaking through a decade of a critical process working on personalities and pressures rather than smart data.

In the spring of 2001, I became CEO of AMIS, a period in which all high-tech and semiconductor companies were suffering major revenue declines.  Yet AMIS was still forecasting, seemingly incorrectly, significant levels of revenue, and were highly confident in their projections. The forecasting process was obviously faulty, yet questions to virtually every member of the management team yielded no insights.  The mantra was “We’ve used this forecasting system for years and it’s worked well.”  Well, it wasn’t working now, as any person one step away from the forest could see.

Accurate forecasting is critical for managing capacity, staffing, capital, work-load balances in the fab facilities, meeting bank covenants and many other forward-oriented decisions. So why didn’t any of the management team have a finger on the flaws? How could they continue to insist that forecasts of $100M+ revenue per quarter were correct in the face of $60-70M per month actual booking rates?  Don’t ask me; I don’t know!  I didn’t ask me. I asked someone else to track specifically how forecasting results were constructed, torturous step by torturous step.  The results were a surprise to everyone.  As it turns out, the procedure had morphed, during the demand-exceeds-supply and growth years, to one in which the individual product units created growth algorithms to inflate the field forecast data. These algorithms were applied by lower level assistants and had lost the visibility of management. Those inflated projections then were used to secure each product department's share of manufacturing capacity. Their inflated forecasts were not really consequential in the prior supply-constrained environment because they tended to cancel each other and were capacity limited.

However, when the market turned down, these forecasts immediately became radically misleading.  Forecasts had been manipulated to the point that they was worse than worthless—they was negatively misleading.  Even in the face of a rapidly declining market, these forecasts meant AMIS was adding personnel, pumping capital into capacity, and expending in other areas. 

Only by tracking the process in detail did we locate the problem. Arithmetically fixing this problem was rather simple once the problem was uncovered. And we took one step toward managing by data, by smart information.