Why your most efficient plant and least efficient plant have never compared notes.

Comparing plants only works once you solve the fairness problem first — otherwise every comparison dies the moment a plant manager objects. Here’s the metric General Mills settled on, and why the same principle works even when your plants don’t make anything alike.
In a multi-plant operation, every facility develops its own way of working. Equipment is different, production demands can vary and local teams solve problems based on what they see every day. That independence keeps plants moving, but it can also make it harder to see which ideas, fixes and energy-saving practices are worth carrying from one site to another.
The knowledge never travels because the measurement never existed.
Leadership may see the value of comparing performance across plants long before the organization has a reliable way to do it. That is usually where the effort stalls. Someone puts two facilities side by side, and a plant manager points out that the comparison is not fair.
Often, they are right.
You can compare commercial buildings using relatively straightforward measures such as energy use per square foot. Manufacturing plants are different. Two facilities within the same company may make entirely different products, use different processes, operate at different scales and consume energy in very different ways.
Without a metric that accounts for those differences, the comparison breaks down before it can produce anything useful. Teams debate the numbers instead of learning from them. And if performance cannot be compared fairly, the practices behind better performance are much less likely to travel from one plant to another.
That makes the measurement challenge more than a preliminary step. Solving it is what makes enterprise-wide learning possible.
How General Mills optimized efficiency across multiple plants.
Twelve years ago, General Mills enrolled in the Department of Energy’s Better Plants program, which asks partners to reduce energy intensity across U.S. operations by roughly 25 percent over ten years. The commitment mattered less than what it forced the company to build.

The company tracks gigajoules consumed per ton of finished product. It is a deliberately unglamorous number, and that is why it holds. It works the same at every plant regardless of what that plant makes or how large it is, which takes the fairness argument off the table and lets the conversation move to why the numbers differ.
The principle transfers even where the specific measure does not. What you need is a denominator tied to what the plant exists to produce — units, cases, linear feet, tons. The test is simple: could a skeptical plant manager raise a legitimate objection to the comparison? If so, the metric is not ready.
Choosing it is half the job. General Mills feeds the data into a quarterly report that compares every plant against its own prior-year performance, and the cadence is what makes it work. A one-time benchmarking exercise produces a slide deck that gets presented once and filed. A quarterly one produces a habit, and it gives plants something to copy from each other rather than something corporate reports at them. A plant manager who sees a peer running better on the same measure usually picks up the phone, because the information is useful to them rather than a judgment about them.
The outlier list is not a budget request.
Once you rank plants on a common measure, the natural assumption is that the high performers need capital. Frequently, they don’t.
Facilities drift in predictable ways, and the most common findings are operational:
- Compressed air. Leaks, pressure set higher than the application requires, air used where a blower or fan would do the same job. The Department of Energy estimates leaks alone can waste 20 to 30 percent of a compressor’s output.
- Controls out of sequence. A setpoint adjusted after a complaint and never revisited. Equipment running through shifts when the line is down.
- Motors on variable loads running at fixed speed. Same energy draw at every demand level.
- Steam and process heat. Failed traps, uninsulated line, condensate that never returns.
Leak repair, pressure adjustment and controls sequencing do not require a capital request. So the first thing an outlier list buys you is not a budget conversation. It is a list of plants where somebody should spend a day with the compressed air system. Larger equipment problems do surface, and when they do, incentives change the math — but leading with operational fixes builds the credibility that funds everything after.
One fix, multiplied.
This is where enterprise visibility pays differently than it does in a commercial portfolio.
At the General Mills plant in Murfreesboro, Tenn., optimizing the compressed air system cut annual energy use by 307,821 kWh with payback in about 14 months, supported by a $36,939 TVA EnergyRight incentive. Compressed air systems look broadly alike across facilities — same failure modes, same pressure creep, same leaks in the same fittings. The real value of that project is not the annual number at one site. It is that the diagnosis is complete, the approach is proven and the business case has already cleared internally. That gives other plants a tested starting point, reduces the time spent proving the opportunity all over again and makes it easier to scale the same improvement across the network.
You do not need to build an energy department.
The objection that closes this conversation is usually headcount. Enterprise energy management sounds like it requires a team most manufacturers cannot justify.
General Mills works with outside firms whose engineers work on-site and are paid a share of the savings actually realized, with no large upfront fee. And then there’s TVA EnergyRight’s Strategic Energy Management program for qualifying industrial customers, which offers free consulting services and engineering capacity across a two-year engagement. Adtran finished two years in the program with 27 identified opportunities, $283,413 in savings, and more than 3 million kWh — more than an internal team already committed to running production could realistically generate.

Take this back to your team.
- Choose one cross-plant metric with a denominator tied to output, and pressure-test it against your most skeptical plant manager before rolling it out.
- Build a recurring report showing each plant its own trend and its position against peers. Quarterly is often enough to build a habit without becoming noise.
- Before adding headcount, look at whether a shared-savings partner or a utility program can supply the same engineering capacity.
The bottom line: The most valuable thing an enterprise view produces is not a ranking. It is a route for a solution found at one plant to reach the ten that need it.
Ready to see what an enterprise-wide view could look like across your facilities? Connect with a TVA Energy Expert today.


