On this page8 sections
Hershey had spent $112 million on that system. Its stock fell more than 8% that day, and the story made the front page of The Wall Street Journal.
I will come back to Hershey in Chapter 4, because the lesson in it is not the one most people expect. Hold one question until then: what was the system missing?
This article is not about which tool is best. It is not a list of apps, and it will not tell you what to buy. It is about why some businesses grow with the tools they buy while others stall with the same ones.
If you love shopping for tools, I may lose you in the next few minutes. Stay anyway. You will still use tools. You will just buy them in a different order.
- The system is king. A tool only speeds up work that already has a shape.
- Name the problem first, then check the four parts, then test small, then keep it up.
- Large companies have lost millions by doing these out of order.
- You can apply all four rules to one job this week.
01The system is king
A system is how your work gets done the same way every time, with or without you. A tool is one part of it.
Think of the tool as the engine and the system as the road. A faster engine on a road that ends in a ditch gets you to the ditch sooner.
Bill Gates made the same point in The Road Ahead, in 1995. Automation applied to an inefficient operation, he wrote, will magnify the inefficiency. It is a striking thing to hear from the man who built a software company.

Jim Collins came at the same question from the other direction. He and his research team set out to learn why some companies leap from ordinary to great. They defined great as investor returns at least three times the stock market over 15 years. They found 11 companies that did it, and they ran 84 interviews with people inside them. Then they asked what drove the change.
About 80% of the executives did not name technology among their top five reasons. The team also found that these companies became technology pioneers only after their breakthrough results, often years later. Collins’s conclusion is that technology speeds up what is already in place. It does not create it.

Read that as an owner. The companies that leapt already had something in place. The tool came later, and it made a good thing faster.
So how do we build a road worth driving on? That leads to four rules.
02Rule one, name the problem before you shop
Before you look at any tool, finish this sentence: "We lose time or money because ______."
Eliyahu Goldratt turned this into a habit of mind. In 2000, with Eli Schragenheim and Carol Ptak, he wrote a business novel called Necessary but Not Sufficient. Its idea is short. Technology can bring benefit, in the words of his organization, "if, and only if, it diminishes a limitation." If it removes no limit, it is a toy.
Then comes the part I find more useful. Long before the technology existed, people built habits, rules and measures to work around the limit. Those habits do not disappear when the limit does. Goldratt asks what benefit any technology can bring if we neglect to change the rules.
Source: Eliyahu Goldratt, Eli Schragenheim and Carol Ptak, Necessary but Not Sufficient, North River Press, 2000. I read the idea as the Goldratt Group describes it in The Reason for Technology, not in the book itself.
Here is the kind of thing to watch for. A web form now sends every new lead straight to an email inbox. Yet the front desk still types each lead into a spreadsheet "just in case." The tool removed the limit. The old habit stayed behind and kept costing the same hour.
How to do it:
- Write one sentence about the problem. Keep it under 20 words.
- Name what it costs you each week, in hours or in missed leads. Use rough numbers, and label them rough.
- Ask what habits grew up around the problem. Those are the rules you may need to change.
03Rule two, check the four parts
Every system has four parts. A tool is one of them.
W. Edwards Deming, the quality pioneer, was speaking to a four-day seminar in Phoenix in February 1993 when he said a bad system will beat a good person every time. It was a remark in a room, not a line from a book. I read it as good news. If your team is not the problem, the setup is, and setups can be changed.

If you already own the tool, keep it. Fill the other three gaps around it first. A tool often starts working once the steps, the owner and the check are in place.
04Rule three, test small before you scale
Here are two true stories. In both, a big company bought new technology, and then something went badly wrong. Read them as a pair and ask one question: what was missing?
Story one: the candy company. Back to Hershey. The new system was a $112 million combination of SAP software for running the business, Siebel software for customers and Manugistics software for the supply chain. It took orders and managed distribution. It went live just as Halloween orders were pouring in.
That is the timing that sank the quarter. CIO.com puts it bluntly: "Hershey’s only real failure was its timing." It also argues that Hershey’s experience was fairly typical. The studies it cites find that most enterprise software installations run late, the business suffers for a time, and revenue can drop for up to six months. The article adds that this kind of software forces a company to change how it operates, so early disruption is normal.
Hershey did recover. In August 2002 it announced a successful upgrade of its SAP software, and one analyst called the 1999 problem a nonissue. In plain words: the tool went live in the busiest season, with no quiet weeks to learn what broke.

Story two: the drug wholesaler. A wholesaler buys medicine in bulk and sells it to drugstores. In 1993, FoxMeyer Drugs was the fourth largest in the United States, with a reported worth of $5 billion, according to CIO Australia. To run faster, it bought an SAP system and an automated warehouse system, and it hired Andersen Consulting to put both in. The project was budgeted at $35 million and scheduled to run within 18 months. The source calls that schedule unrealistic.
Three existing warehouses were closed. Workers whose jobs were threatened by the automation opposed the project, and the first automated warehouse suffered sabotage, with damaged inventory and unfilled orders. The source says the new system handled 10,000 orders a night, compared with 420,000 under the old mainframe. FoxMeyer went bankrupt in 1996 and was later sold to a competitor for $80 million. In plain words: the new setup could not handle the real workload, and the people it affected were not on board.
Source: CIO Australia, IT’s biggest project failures. A press account of the case.
Why did FoxMeyer fail? People still argue about it. In 1998 its bankruptcy trustee sued the software maker and the consulting firm, for $500 million each, which suggests it blamed them. Those suits were settled or dismissed in 2004. But Christopher Cole, an executive at the warehouse vendor, said it was not the technology’s fault, though he had a stake in the answer. He said it was a management failure. I can’t tell you who is right.
Here is what I can say. Put the four parts from Chapter 3 next to the reports. At FoxMeyer, people opposed the change, and the new system did not match the volume of the old one. That is a People gap and a Tools gap before anyone had a chance to build a check. At Hershey, the setup went live with no room to find breaks.
From the reports I found, neither company started with one small job. Both tried the new setup at full size. Here is my rule for now: try any new setup on one job, for a few weeks, before you move anything else onto it. Ask the owner to write down what breaks. This is my suggestion, not a research result.
05Rule four, keep it up
The best example I found of a system that lasts is Toyota’s. Long before the factory system, Sakichi Toyoda (1867 to 1930) built automatic looms that stopped on their own when a thread broke, according to the Lean Enterprise Institute. That is a check built into the machine. Taiichi Ohno, who led production at Toyota after World War II, started the Toyota Production System in machining and led its development through the 1950s and 1960s. It spread to suppliers through the 1960s and 1970s.
What kept it alive? The Institute’s answer is plain. It says the system "is maintained and improved through iterations of standardized work and kaizen." Kaizen means steady, small improvement.
Source: Lean Enterprise Institute, Toyota Production System.
That is the lesson I most want you to take. A system is never finished. In a small business, your version of the self-stopping loom is a weekly look at the work. Name an owner and a backup. Pick a day each week or month to check it. Ask three questions: Is it still doing its job? Did our work change? What is the next fix?
I go further into this in Every system needs an owner.
Where tools do help
I don’t want to overstate this. Tools are not the enemy. A good tool on a well-built system can save hours. Collins himself said technology can accelerate a company that is already moving.
One honest caution about the evidence. Large companies are not small businesses. Boston Consulting Group reports that 70% of digital transformations fall short of their objectives. Its research draws on a survey of 825 senior executives. BCG says "fall short," not "fail." It shows a pattern. It does not prove a law.

Your five questions
Can’t answer three of them? You have found your real project.
By Angel Flores


