5 Reasons Why Store Checks Aren't Accurate

| About: J.C. Penney (JCP)

Individual store checks get mentioned fairly often in comments and on message boards as evidence that a department store is doing well or doing poorly. However, these store checks are actually a very inaccurate method of estimating department store performance. Too often, reports of busy local stores from a department store chain are used as an argument for why a stock is worthy of a buy, while reports of quiet local stores are used as an argument for shorting a stock.

Individual store checks should actually be very far down on the list of items influencing investment decision criteria. I will primarily be using examples from J.C. Penney (NYSE:JCP) since I am most familiar with that company, but will also discuss Macy's (NYSE:M) and Kohl's (NYSE:KSS).

Here are five reasons why you should ignore individual store checks in an investment decision.

Small Sample Size - Number Of Stores

J.C. Penney has approximately 1,100 stores. Kohl's has around 1,150 stores. Macy's has around 850 stores. If you live in a major city and have some time, you may be able to check out 6-7 nearby stores from each chain.

Although local store performance is significantly influenced by corporate initiatives, it still is not possible to accurate gauge nationwide performance based on a sample that is geographically very localized and represents less than 1% of total stores. Even if you knew for certain that there was a 10% increase in sales for those six stores, that would only result in a sales increase of less than 0.1% for the whole chain. The value of a 0.1% increase in sales can be measured in pennies per share at best.

A very rough analogy would involve the weather. One city may have a very cold winter, but from looking at that one city there is no way to tell if the average winter temperature for the U.S. is colder or warmer than typical.

Small Sample Size - Hours

Traffic varies significantly depending on when you visit the store. One study estimated that 11.6% of Americans are out shopping on weekends between 2pm and 4pm. This decreases to 4.1% on weekdays between 6pm and 8pm. So if you visit a store at 7pm on Wednesday and then on Saturday at 3pm, you'd see 183% more people on average on Saturday without anything else changing.

Similar differences may also happen depending on time of year. The back-to-school shopping season and the holiday shopping season are notable for their busyness.

Very few people would be able to accurately tell you how busy a particular day and time is supposed to be in relation to another day and time. Without being able to adjust for that, a store may seem very busy at one time and very quiet at another time just based on normal traffic variation.

As well, checking out a store for a couple hours gives a very small sample size even when trying to estimate monthly sales for that store. A random J.C. Penney store like its Alderwood Mall location is open for over 300 hours each month and over 900 hours each quarter.

Lack Of Baseline For Comparison

One of the more influential numbers in retail is change in same-store sales. This number is generally useful because of the context that it includes. Sales this year are compared to sales last year over the same period for the same stores. We are able to make a proper comparison as a result.

Doing a retail store check on the other hand does not typically include any sort of baseline. Without knowing what sales were last year for that store during the same period, we cannot put any of the store check numbers in context. Is 500 customers in the store good or bad? As noted above, it really depends on the time of the visit and the particular store in question. Even if you do figure that out, it is still only data for one store over a short time period.

Difficulties In Estimating Actual Traffic Levels

Let's look at a 100,000 square foot department store that does $200 of sales per square foot per year. That results in $20 million of revenue per year. This would represent a stronger location as J.C. Penney averaged around $116 per square foot in 2012.

If the average transaction is $75 and 20% of visitors purchase something (conversion rate is in the low-20s for J.C. Penney), that means that there are approximately 1.33 million visitors to that store each year.

Assuming the store is open 75 hours per week, there are roughly 340 shoppers visiting in an average hour in an average week.

The difference between a good quarter and a poor quarter may only be 5% or less. Macy's was applauded for same-store growth of 3.5% during its latest earnings report vs. expectations for a 2.1% increase, while Kohl's took a hit for announcing same-store declines of 1.6% vs. analyst expectations for an increase of 0.7%. Can anyone easily tell the difference between 335 shoppers visiting the store in a hour and 350 shoppers visiting the store in an hour without counting everyone coming through the entrances? That's what a 5% difference represents, yet that sort of difference resulted in Macy's gaining nearly $2 billion in market capitalization and Kohl's losing nearly $1 billion.

As well, many of the reports of store visits I have read do not even give numbers. Busy or quiet may mean something significantly different in terms of numbers to one person compared to another person.

Asking Store Employees Doesn't Give An Accurate Picture

Earlier this year, Matt James collected quotes from J.C. Penney store leaders about 2012 holiday sales. He combed through news sources and found 16 positive quotes from store leaders mentioning that the stores were busy and that sales exceeded expectations, and were comparable or better than last year. He did not find any quotes during his search that hinted at poor or mediocre sales.

Despite the positive picture from store leaders, the reality was that J.C. Penney had an extremely poor holiday season in 2012. Same-store sales decreased by 31.7% in Q4 2012, which is probably one of the worst declines in department store history. What this tells us is that store leaders and employees are going to be very positive about sales results no matter how the store is actually doing. They may be a bit more candid off-record, but not if they think that what they say is going to be shared publicly. That's also assuming that the particular employee has an accurate picture of sales trends.


Individual store checks may be useful in getting a feel for a department store's merchandising and pricing strategy. However, they should not be used as a gauge of performance for an entire chain. Too often I read about how a particular store is very busy or very quiet, and that gets used as part of an argument for why the department store chain will either beat or miss its sales numbers for the quarter. Given the problems with these store checks that I outlined above, visiting your local stores to gauge traffic should not be something that you use to influence your investment decisions in any meaningful manner.

For store checks to have any validity, there would need to be checks of dozens of stores from across the country. These checks would likely need to cover a period of many days, not hours, and would ideally need to be compared against a baseline of the same stores over a comparable period last year. Information from store managers or employees has not been proven to help unless they are willing to go on record with actual numbers.

When making an investment decision, I would suggest only paying attention to checks that have come from a broad-based check of many stores over a lengthy period of time. These checks need to also provide numbers (not qualitative data like busy or empty) and be compared to a baseline involving a similar period for the same stores last year. Such information tends to be hard to come by, so I typically only pay attention to traffic and sales data provided by the company itself.

Disclosure: I am short JCP. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.