Warranty Marksmanship:

Like target practice at a summer carnival, companies take careful aim with their warranty expenditures, deciding how much to set aside to pay future claims, and how much of a financial cushion they think they need to keep in their warranty reserve fund.

There's no such thing as an ideal level for either warranty accruals or warranty reserves, but after studying some 900 U.S.-based warranty providers for the past 12 quarters, we've calculated that the average accrual rate is 1.8% of product revenue and the average warranty reserve balance is equal to what a given company pays out in claims during a 17.2-month period.

Very few companies are at that level and why would they want to be? After all, it's just a weighted average of all manufacturers. In last week's newsletter, we showed how different industries behave differently. Some experience relatively high levels of claims, and so they make relatively large accruals and keep relatively large reserves. Others experience below-average claims rates, and so they make relatively small accruals and keep relatively small warranty reserves. And still others mix large and small, for instance by keeping larger reserves than average despite a low claims rate, or by keeping smaller reserves than average despite persistently high claims rates.

On Top of Their Game

We suspect that the companies who are the most on top of their game show the least fluctuation in either accrual rate or warranty reserve. In these charts that compare reserves to accruals, such companies would be represented by 12 dots in a tight cluster, as if a marksman had taken aim at the target and landed every shot. Take them to the carnival and they'll win the biggest plush toy.

Conversely, companies who are the least in control of their warranty process would have the worst aim, scattering their shots wildly, if they land at all. One quarter, accruals might be 2% and another they'd be 4%. The "thickness" of the warranty reserve cushion would rise and fall accordingly.

Last week, we plotted reserves and accruals for 12 different industry sectors, and found that each returned a somewhat tight cluster of dots. We plotted only the five largest warranty providers individually (GM, Ford, HP, Dell, and IBM), and found that they too returned a somewhat tight cluster of dots. One could suggest that maybe this is how it is for all companies. But that is definitely not the case.

This week, we'll take a closer look at the building trades, to illustrate how in and out of control of their process some warranty providers can be. Warranty Week tracks some 157 companies who report warranty expenses in the home construction and furnishing sector. That includes not only the builders of the actual homes, but also the makers of the tools and machinery that builds the homes, as well as the heating and cooling systems, appliances, furniture and fixtures that fill the homes.

As was done in the automotive and computer industries in recent newsletters, this week we divided the entire building trade sector into smaller groups, to see if companies with similar products have similar warranty experiences. As usual, we oversimplified, chopping the group into 1) homebuilders, 2) HVAC, 3) appliances, 4) building materials, and 5) tools and machinery.

No OEM Effect in Homebuilding?

There's been a lot of offline discussion about this topic. One reader asked if the same effect seen in the automotive and computer industries -- where the name brands sold to the end user customer end up paying most of the warranty expense -- was also present in other industries. After completing this study of the building trades, we can say conclusively that it's not present here. Unlike the GMs and HPs of the world, the builders get off relatively lightly when it comes to warranty claims.

In response, the reader suggested that's probably because the builders -- the "OEMs" of this market sector -- don't have as strong of a brand name as do Ford and IBM. Another reader said the builders are tougher on their subcontractors than the automakers and computer OEMs are on their suppliers. And another reader said the homebuilders report lower levels of warranty spending simply because relatively few incidents rise to the level where money has to change hands. Instead, somebody stops by, "off the books," as it were, and informally fixes whatever needs to be fixed.

No matter. Even if we don't know why, it's clear from the charts that follow that the appliance manufacturers (Figure 3) make the largest accruals, and that the HVAC manufacturers (Figure 2) keep the highest levels of reserves. The "OEMs" make the least accruals, though a few do seem to keep relatively high levels of reserves.

As seen below in Figure 1, the ten largest homebuilders are for the most part to the left side of the chart, meaning they're reporting warranty accruals below the 1.8% average observed for all warranty providers. The notable exceptions are Fleetwood Enterprises Inc. and Champion Enterprises Inc., which are both in the lower right quarter of the chart.

It turns out that of these ten, Fleetwood and Champion are the only two that manufacture mobile homes and/or prefabricated homes. The other eight sell site-built homes. We think there must be a fundamental difference between the two, if not in terms of the product then certainly in terms of the warranty process. Even glancing at Figure 1, it's pretty clear that the eight are different from the two. So let's redraw the chart with only the eight makers of site-built homes, and call it Figure 1a.

In order to use a vertical scale that goes from 0 to 40 months, and a horizontal scale that goes from 0% to 2.5%, we had to ignore one of the 96 dots. Beazer Homes has one outlying data point for accrual rate in the first quarter of 2005. During that quarter, because of mold and moisture intrusion problems and the likelihood of future warranty claims, the company put aside 5.6% of its revenue -- almost $54 million -- and expanded its reserve fund to nearly $137 million in size.

That dot appears all by itself in the upper right corner of Figure 1. It does not appear in Figure 1a. Although Beazer Homes doesn't show the same kind of tight clustering as Pulte -- and indeed it's the only major homebuilder to the right of the dotted line that represents the average -- it's pretty obvious even in Figure 1 that this particular dot is the warranty equivalent of a stray shot.

The warranty planners and forecasters of the other seven companies, with the possible exception of M.D.C. Holdings Inc., display their skill by turning in a relatively tight grouping of dots. It doesn't matter if there's mixing or overlap between companies. In fact, a few charts down, the colors mix like a Jackson Pollock painting, which we take to mean that similar companies behave similarly.

Leading HVAC Manufacturers

In Figure 2, we charted ten of the largest HVAC manufacturers, keeping in mind that Carrier is only a part of United Technologies and Trane is only a part of American Standard. Perhaps that's why they're a little to the right and above most of the others? Perhaps the diversity of their product lines causes them to keep a slightly thicker financial cushion on hand when it comes to warranty reserves?

By the way, we apologize in advance if the labels aren't always near the dots. In some cases, as in Figure 2, there's such a crowding effect that there's not enough room for all the company names. However, they're all color coded, so when in doubt just look for the label that's the same color as the dots.

We should note that United Technologies had an outlying data point in the third quarter of 2003 that doesn't fit onto this chart. In addition, Ingersoll-Rand had an outlying data point in the third quarter of 2004 that wouldn't fit. In both instances, the capacity of their warranty reserves soared after claims temporarily plummeted. During the other 11 quarters, their warranty experience was much more well-behaved.

In general, the 10 companies charted in Figure 2 had higher reserve levels than the eight companies charted in Figure 1a, though most seemed to have similar accrual rates in the range of 0.5% to 1.5% of product revenue. This means that although both HVAC manufacturers and homebuilders put aside the same percentage of revenue, the homebuilders feel comfortable with a thinner financial cushion. While the homebuilders are comfortable keeping an amount on hand equal to what they pay out in 12 to 16 months, the HVAC manufacturers fell more comfortable in the 20 to 24 month range.

Longer Warranties = Larger Reserves?

A reader asked if perhaps there's a correlation with warranty duration. That may be part of the answer, with HVAC systems typically featuring some of the longest warranties in the business. Therefore, a thicker reserve cushion may be more appropriate than if the warranty durations were all only 90 days. And within the sector, it may explain why some of the component manufacturers like Rockwell Automation and Honeywell International are at the bottom left while many of the complete system manufacturers are at the top.

But how can we explain the noticeable difference between Figures 2 and 3? The companies in the chart above are makers of the big central heating and air conditioning units and their components. The companies in the chart below are making window air conditioners, refrigerators, washing machines, clothes dryers, ovens, toasters, and numerous other types of household and personal appliances. One would expect they'd show similar diffusion patterns for reserves and accruals. Clearly, that's not the case.

We should note that during 2003 and 2004, TurboChef Technologies Inc. missed the chart completely on multiple occasions -- on the high side for both warranty reserve capacity and for accrual rate. As a result, only five out of 12 dots fit within this scale. As we detailed in an article earlier this year, they're the makers of the popular sandwich-heating ovens used by the Subway restaurant chain. Here, the scatter-shot pattern is confirmation of the suspicion that the runaway sales success of the ovens and the heavy workloads they experienced caused TurboChef to lose control of their warranty process. In fact, most of the chaos took place in the second half of 2003 and the first half of 2004. In the last few quarters, it seems as if TurboChef has regained some control of its warranty process. So now at least it's hitting the chart, even if its aim needs improvement.

Appliance Industry Average

Note also that Maytag, Whirlpool, Fedders, Helen of Troy, and Jarden, and the countless appliance brands they represent, are in their own cluster around the 3% accrual and 5-10 month reserve levels. We'd venture to say that this must be something of a benchmark for the appliance industry. It's NACCO, Ametek, and Applica who stand out from the crowd -- the latter for its unusually high accrual rate, and the former two for their lower-than-average accrual rates. Anything very far outside that 2% to 4% range seems to be unusual for appliance makers.

In the automotive and computer industries, the OEMs had the highest accrual rates and thickest reserve cushions, while the suppliers generally had the lowest and the thinnest. Somewhere in the middle were the drivetrain manufacturers and the makers of disk drives. If there was any analog in the building trades, the HVAC and appliance makers would be the drivetrain people in the middle. However, by their positions relative to the dotted-line averages, it's clear that there's something fundamentally different about HVAC and appliance manufacturers.

At the same time, there's something fundamentally similar about the homebuilders of Figure 1a and the makers in Figure 4 of all the fixtures, furniture, and building materials that go into the typical home. Both groups seem to get off relatively lightly when it comes to warranty expenditures. That includes all the electrical and plumbing fixtures, as well as all the cabinets, carpets, windows, doors, flooring, and other warranted products.

So while in automotive and computer, it's the customer-facing brand names that end up paying most of the warranty expense, in the building trades it's the HVAC and appliance makers who end up holding the bag. Perhaps that also has something to do with the visibility of their brand names? What we mean is this: you know who built your car, and who built your computer. And you probably could guess who made your refrigerator and the other major appliances.

Now, who built your home? What brand of carpet do you have on the floors? Who made the countertop in the kitchen? The windows? The cabinets? What brand of furniture did you buy? Only heard of a few of the top ten manufacturers in Figure 4? Well perhaps that's the point. We've all heard of Whirlpool and Carrier, but who is Masco? By the way, BMCA is our abbreviation for the Building Materials Corp. of America, makers of doors and roofing materials.

Finally, with Figure 5 we had to create a chart for all the construction tools and equipment that is used to build the home, not to mention all the outdoor power equipment that keeps the home looking good. And so, as with peripherals in the computer sector, this is the "other" category that includes everything from cranes and earth movers to lawn mowers and hedge trimmers.

We can't in all honesty say there's much of a pattern here, although it's notable that Toro, John Deere, and Caterpillar seem to have their own cluster on the right side, while Textron , Black & Decker, Briggs & Stratton, Manitowoc, and Terex have another cluster in the middle. Stanley Works, Black & Decker, and Terex each have their own tight clusters, while Foster Wheeler Ltd. is such a scatter shot that only 3 out of its 12 dots fit on the chart.

Warranty Marksmanship

Remember, each of the 50 companies included in these charts is capable of deciding how much to put aside in warranty accruals and how much of a cushion to keep in reserve. Somebody decides how much is enough. If they guess too high or too low, sooner or later they have to correct themselves. And after 12 quarters of poor aim, they'll look like Foster Wheeler or TurboChef, with numerous shots not even hitting the target.

However, if they are always close to right, if they anticipate the right level of claims, and if their products have consistent and predictable levels of quality, then they'll earn their warranty marksman pin. After 12 quarters of good aim, they'll look like Black & Decker, NACCO, Textron, or Pulte, where the dots are so close together that it doesn't look like their are 12 of them.