Franchisors keep telling me stories of people trying to sell profiles by doing a "benchmark" and thereby "customizing" the profile for the franchisor. Their method? It sounds so easy. Just take 5, 10 or 20 of your top franchisees and profile them. This will supposedly give you a "benchmark" unique to your business.

Sounds reasonable doesn't it? What a load of hog-wash! There is no way that you can arrive at an accurate assessment of a profile by profiling only 5, 10 or 20 people. Especially when you only profile people in one performance catagory. Why? First let me explain exactly what a "benchmark" is. Then I'll give you just a couple of reasons.

A benchmark, by definition, will tell you nothing about differentiating between performance levels. The definition of a benchmark is simply - "A standard by which something can be measured or judged". Sure, now you have something to compare franchise candidates to. But, and it's a big but, How do you know that whatever tool you're using for the benchmark actually measures the characteristics that make good franchisees good? Do you take it as an article of faith that the tool measures what needs to be measured? Why would you? If you are taking it on faith, then why would you even bother doing the benchmark? Without profiling poorer performers, you'll never know if the profile tells you anything useful.

Let's get back to the definition for a minute. Notice that nowhere in the definition does it mention comparing that standard to performance? But isn't predicting performance the exact reason why you're considering profiling your candidates?

What profile buyers are really looking for is: "A standard that, when used to compare individuals, can accurately differentiate between performance levels. " A big difference. What salespeople are trying to tell you they can do by profiling only 5, 10 or 20 people with a "benchmark" is demonstrate that the profile results have a direct relationship with performance without measuring performance. Kind of difficult to do, I think.

If you really want to find out how profile scores relate to performance, the profile has to be validated. By that I mean that franchisees across a range of performance categories have to participate and performance data on those franchisees gathered. Then the performance data is compared to the profile scores. The purpose is to see if the profile scores alone can differentiate between the performance categories.

Properly designed validation studies can be invaluable. With them, you can see if the selection system, and the profiling tool being used, are working as well as they could. Oh, did I mention that to be statistically significant, you need to include at least 100 franchisees in the study? Anything less and you could simply be picking up "noise" in the data.

Let's look at some other issues with the proposed method of profiling 5, 10 or 20 people to arrive at an "ideal" franchisee score. If you don't include your worst performers, what have you learned? Very little. You know that the people you profiled are good performers. But you knew that already. And you know what their profiles look like. So what? How do you know what a poor performer's profile will look like? How will you know what the differences are? And on top of that, without including poor performers, how will you know if you're even measuring the characteristics that differentiate performance? Never mind whether the profile can actually predict performance! You've heard the expression GIGO (Garbage In, Garbage Out). Here's a perfect case for it in action. Too small a sample size and no one to compare good performers to gives you garbage.

In fact, you could develop an equally accurate benchmark without going through the bother of profiling anyone. Just place upper and lower limits on the profile dimensions being used. You will still have "A standard by which something can be measured or judged". Much simpler. Much less hassle for you and your franchisees and, in the short run, much cheaper. Plus, it'll be just as accurate. Because you already have a good handle on the characteristics of a successful franchisee.

Without knowing what a below average performer's profile looks like, how can you make a reasonable assessment of who will perform at high levels? And again, maybe the difference between good and poor performers isn't even assessed in the profile being used! Maybe the profiles of good performers and bad performers are very slight and are limited to only one dimension. You'll never know if you don't include both the good and the bad.

If you want an accurate and meaningful study, people from different performance groups absolutely must be included. When we do studies, I recommend that about 50% of the franchisees included are performing at average levels, 25% at below average, and 25% at above average levels. The more people you include, the more accurate the results. We always resommend a sample size of at least 100. Because we want to be able to say, with confidence, "This dimension on our profile translates to $X in annual sales."

It's the same with the offers made to do "one free profile". If you're considering whether the profile can help you decide whether a candidate would make a good franchisee, what good does one profile make? Unless you profile a good performer and a poor performer, how can you in good conscience say that the profile can accurately differentiate between performance levels? The truth is, you can't. Simple as that.

And that's the whole point of choosing to use a profiling system, isn't it? To be able to tell which franchise candidate will perform at a high level.

Here's another reason. Let's say you profile 20 people. Big deal. Ask any person with a knowledge of statistics what kind of results you'll get from that. Their response? Not much... if anything. A sample size of 20 people is simply too small to give meaningful results. Even if you include below average performers in the group.

In any statistical analysis, the larger the sample size, the more accurate the results. Because our sole purpose is predicting performance of franchisees, I always recommend the largest sample size possible. Preferably over 100. While studies with a minimum sample size of fifty are possible, the results are far less statistically accurate.

The more people included in the study, the more performance categories can be used. For instance, if you only include 50 people in the study, don't expect significant results from more than three, and possibly as few as two, performance ratings. With groups over 100, knock yourself out. Define as many different performance categories as you like. Running an analysis on ten groups is almost as easy as running it on three.

But what if you don't have more than 50 franchisees operational? Then you make sure that the profile you are considering has been shown to be consistently reliable and accurate in predicting performance of franchisees specifically. If the profile has demonstrated reliability and accuracy, there is no need to even consider trying a benchmark with such a small sample size.

Instead, do your own a benchmark by placing upper and lower limits on the profile dimensions being used. You'll be in the ballpark. You wouldn't have survived this long without knowing already, would you? And, it will be just as accurate as profiling only a few of your best franchisees.

So you see, a real validation is much more than simply profiling a few of your top performers. If you want to make a decision "by guess and by golly", then go ahead. Just be prepared for the fall-out. Because the fall-out will happen. I guarantee it. But then, if you are willing to make your decision "by guess and by golly", why would you be looking at adding profiling into your selection process?

By somehow believing you can get a good idea of what to look for in a franchise candidate by using such a small sample size and not including performance data, you are wasting your time, money and the candidate's money and time. One thing we all need to keep in mind.... our decisions have huge impacts on the lives of the franchise candidate and their family.

It's incumbent on everyone involved in franchisee selection to make sure that we do all we can to help everyone be successful. Basing a decision on flawed assumptions only costs everyone. Usually big-time. There are no small costs associated with franchisee failures.

Properly used, profiles can play a critical role in selection. But basing any hiring decision or approval for a franchise on a flawed "benchmark" simply gives profiles of all kinds a bad name.

To sum up, profiling only your good performers (no matter how many) doesn't:

1. give you a statistically accurate analysis of the characteristics of the group.
2. It doesn't even give you a meaningful basis to describe your culture statistically.
3. It doesn't identify attributes and characteristics of people at different performance levels.
4. And you still won't know if you're measuring the characteristics that differentiate performance!

For that matter, it gives you none of the information a valid study should, and would, give you.

Profiling only your good franchisees does:

1. Give you a false sense of security thinking you've identified your "ideal" franchisee.
2. Describe the people you've profiled. But is that all what you're looking for?
3. Give you a chance to waste your franchisee's (and your) time on a flawed experiment.

One of the main reasons that top franchise organizations succeed is because they recruit and retain the right people for their franchise opportunity. The best franchisors look for ways to eliminate, or at least reduce, personal bias and subjective criteria from the selection process.

A well-designed validation study:

• provides statistical analysis of the attributes and characteristics of your franchisees,
• provides a meaningful basis to describe your company's culture statistically,
• provides an ideal candidate model to fit your franchise,
• provides a database to which future franchise candidates can be compared,
• identifies attributes and characteristics of franchisees at different performance levels,
• indicates how well the instrument has been working, and
• provides a real basis by which to evaluate and refine the franchisee selection system.

Validation studies are invaluable because with them, an organization can determine if its selection system is working as well as it could. And if not, find possible ways of improving the selection process.

A note of caution. Study results are alway dependent on:

• The number of franchisees participating in the study.
• The distribution of the franchisees in the study. How many franchises are rated as Above Average, Average or Below Average.
• The amount and quality of the objective data provided.
• The quality of the subjective data provided.

Like the old saying goes: “Garbage In, Garbage Out’. So for the most accurate results, it´s crucial to be brutally honest in rating the franchisees.

What to Consider For A Valid Study
So, how do you go about getting meaningful results? First, we need to decide on what performance criteria to include. Basically, the performance criteria should be those factors which define “Success’ as you see it. Typically, the definition of success includes both objective and subjective criteria.

Always Include Objective Performance Data
The performance data can be subjective or objective data or a combination of both. Objective data is valuable because it often reflects actual performance with little or no personal bias. Any performance criteria that can be measured in a neutral manner would qualify.

Typically, objective performance data includes such things as “Actual Sales Volume’, “Staff Turnover Rates’, “Profitability’ and so on. Studies can also be broken down by region, number of locations each franchisee operates, whether the franchisee is an owner-operator, an investor, a multi-unit operator, or other specific criteria that are significant and important to your organization.

Consider Subjective Data
Subjective data is useful because it reflects the management evaluation of the person. However, it can be biased by the nature of the personal relationship between the people involved. The rater's perception can occasionally affect the rating and the actual performance of the franchisee may be affected as well.

Subjective criteria are things like ratings on performance like as “Relationship to Head Office’, “Employee Morale’, whether the franchisee “Meets Location's Sales Potential’, “Overall Rating’ and similar issues.

Biographical data like: education, the recruiting source and other test results, can also be included to determine if there is any correlation between them and performance in the franchise. Although the use of biographical data is not critical, it can provide useful information in determining which factors are proving helpful in your selection system.

And, to make sure the profile is EEOC compliant, you'll need to track the race, sex, and ethnic composition of each franchisee.   

There are many possible performance criteria that can be used. If you are uncertain about what to include, we will work with your organization to select the criteria that will provide you with the most relevant results.

The Importance of Sample Size
The next decision is whether you want to include all, or only some of your franchisees in the study. In any statistical analysis, the greater the sample size, the more accurate the results.

While a study with a minimum sample size of fifty is possible, the results of a study of this size are far less statistically accurate. If you have less than 100 franchisees in your system, I strongly recommend all franchisees be included.

There are some very good reasons for choosing to invite all franchisees to participate in the study.

1. There is no cost difference over including all franchisees vs. only some of the franchisees.
2. By inviting all franchisees to participate, you get a much better cross section of franchisees in all performance ranges. For the study to be valid, franchisees from all performance categories must be included. Anything else is a waste of your time and money.
3. The more franchisees that participate, the more accurate the results.
4. The scores of every franchisee does not have to be included in the study. We can select those you want to include. For instance, if you have 500 franchisees and invite them all to participate, only 200 may respond. Then, of those 200, you may choose to only include 150 of them for various reasons.

    

Plus, inviting all franchisees to participate eliminates the old franchisee paranoia issue. You know the one. It when a franchisee that wasn't invited asks: "What's wrong with me? How come you invited XYZ but not me? Aren't I good enough?"

When we do studies, if we can't include everyone, I generally recommend that about 50% of the franchisees are “Average’ performers, 25% “Below Average’, and 25% “Above Average’ performers.

Using the methods listed here, you'll know:

1. You have an unbiased, accurate assessment of what it takes to be a good franchisee in your system.
2. How your top performing franchisees differ from your poor performing franchisees.
3. If your franchisee selection system is working at peak efficiency, and if not, how you can improve it.
4. If the profile actually measures the things that predict performance.

And that's how you go about establishing a valid benchmark. One that will help you target the best franchise candidates.

Dynamic Performance Systems is a member of the International Franchise Association Supplier Forum.