The first indicator is usually stats. You can take a stat book of an org and look over its GDSes and know their interrelationship and find the outpoint, and then from that outpoint you will know what part of the org's folder to read. If you are doing evaluations by reading the whole folder, you're being silly. You're not interested in that. You're interested in this outpoint, because that's your first outpoint. Your first outpoint usually occurs in stats.

One outpoint, from stats, was tremendous quantities of bulk mail being mailed at vast cost after the stats had been brought up by regging, and then the stats collapse. That was the first oddity that was noticed from some Dissem stats. So it was a stat oddity. They were busy regging and they made a lot of money, and then they spent it on bulk mail and went broke. Because there was a stat oddity here. It meant the GI did not match the bulk mail. So it's an outpoint. It's inconsistent. Contradictory. Something's false. So right there, you're looking at a great big cracking outpoint. One or the other of those facts is a lie, or something's wrong. And we find out the real outpoint underlying it is wrong target. It's just number of pieces being sent out. They were mailing out fliers several times a week-sending scraps and calling it bulk mail.

Now just the fact that an org's stats are down is an outpoint.

Having found a downstat you look to see if the org ever did make money? If it was ever affluent. Just taking it from the standpoint of GI, was this org ever affluent? If the org was ever affluent, it must have been doing something right so you've got something that approximates its ideal scene.

You haven't approached data files yet. That's why stats are separate from the data files.

So here's two conditions: (1) the stats are down, and (2) you can't evaluate one thing, as you learn in the Data Series, unless you have a comparative thing. You have to compare it with something. So you can find a period when their stats were up.

You find out that in July of 1969 Kokomo was really booming. It had nice climbing stats and they went up and up and up and up and up. And that rise started on the 6th of June. What did they do? In May and June of 69? Those are the two folders you want. Anything you can find out about that org of May/June 69. That gives you something dimly resembling an ideal scene. It isn't the ideal scene, but it is certainly an upstat scene. That gives you a comparative.

If you were hot you would use your telex lines to fill in the missing holes. For instance, if you don't understand something, or if it looked like they moved in 1970 and you can't find out locally, and you don't seem to know whether or not they did­location seems to be something important here-you could send a telex to somebody who might know and say, "Where were you located in June of 69? Where was this org located? Can you find out from anybody?" It might be important you see. This is just a collection of a little bit more data. You know that the org was doing something, at that time, that it isn't doing now.

I did just this when I wrote the PL "Selling and Delivering Auditing." I looked back when HGCs were really making the money and wrote that PL. This PL is in use in one org and they're really going to town. They're using the same system. A guy comes in to sign up, they say, "No you can't sign up for one intensive, thank you, you'll have to buy seven," or something. So he does, he pays the money on the barrelhead. That PL comes out of a comparative-a comparative of HGCs not selling much auditing and having a hard time doing so, and what they were doing in an earlier period.

So, when doing an evaluation (1) look at your stats, (2) find your outpoint in the stats, (3) find some comparative-find some period of affluence for the org, if you can, to give you some ideal scene for that org. That requires something of a pluspoint evaluation. Now you can do your outpoint evaluation. Because you've already got the outpoint, you don't have to read 8,752 folders.

A while back, I asked the Data Bureau for the folders of a particular downstat org. The first folder came up, that wasn't even a complete month's folder. I looked through the folder, read scraps of what I was reading, picked out the reports I wanted. Scanned them. Pulled the outpoints out of them. Counted up the outpoints as to where they were going. And the thing just fell apart. The CO was unaware of the fact that Personnel was letting him down. That was their admin Why. And obviously the CO had to take that person in there off. And obviously there was something wrong with this CO. Now every eval done on that org since is grooving on straight down that same Why. We've tried to make orders, and we've tried to do this and we've tried to do that. But now an ethics situation has developed out of the thing. We got the admin Why all right. But an ethics situation developed as we tried to get this in. And notice that the ethics situation develops when you try to get in the Admin or Tech why.

In another area the ethics situation developed to such a degree that it then emerged-after an observation mission, after a handling was done and orders were issued-that they did not execute a single one of them. They were told to revert. They did not. Therefore an ethics Why was looked for.

Now I've just found out why people can't put in ethics. They don't know investigatory tech, and possibly in some cases their own ethics are out. If you put their own ethics in, they will get in ethics further. The reason they assign broad conditions and the reason there are so many Comm Evs is they don't know how to investigate.

Someone was given an evaluation to do and had been on that for five days. I kept asking all this time-where's this evaluation? People must think I'm rushing them. Evaluators are slow because the evaluation is not being done in this sequence: (1) stats, (2) who was on where.

I gave an order to an evaluator to find out exactly when did a CO of an org come to Flag, and when did this person go back, because that would give you a stat comparison. That was how I found this person was the man-of-all-work and the scooting genius of that org. Now you're talking about ethics. It's the police action called date coincidence. It's how you locate geniuses and murderers. Body found in swamp. Her cousin arrived in town on Tuesday. Body found on Wednesday. Guy departed on Thursday. That's all the police need. That's called date coincidence. That's old time investigatory tech. It's still with us.

So, when were they gone out of the org, and when did they arrive back in the org, and'what happened during that period of time? Important!

In the case of this particular CO, I found out that two other execs could leave the org and return and nothing happened-but when the CO left, the roof fell in, the front steps collapsed under everybody, and the staff went on vacation. I traced this down and I found out that this CO would run around the org wearing hats in rotation. She dived into Tech and wore the Tech Sec hat for a while, and then she dived into another area, and she wore that hat for a while, and the stats would go up. In other words, she supported that area by punching one area at a time. That was the way she was operating. So if she was all over the org like that, her obvious post was D/CO. We put her on that post, and the org has done well ever since.

Now that's a sort of ethics action in reverse. That's looking for who really pushes it. You don't just keep on looking for tigers. Tigers are probably more numerous than geniuses. But you could find that certain people have a vast effect on stats. This is how you evaluate a personnel scene. In another org, a guy took over and the place has been crashed ever since and it was right square on the stats. There is your most obvious ethics investigation by stats.

When you don't know, you've got to send an investigatory mission and it's got to be run well. Otherwise they just wind up shooting all the people that the staff complain about.

If you don't operate on a comparison every time-comparison admin Why, comparison on the stats, ethics comparisons-if you're trying to operate on a single datum, that single datum won't buy you any pie. Because it has nothing to compare with.

What the Data Bureau gives us is experience. And that is huge files full of experience, but you've got to recognize what you're reading. You don't read everything! If you do you're omitting an analysis of the GDSes and an analysis of who went on where. At a good time and a bad time.

What are you looking for? You're looking for the stat-look at your GDSes (this is for your admin Whys), tells you the big outpoint, tells you what information you're looking for in the files-and you're only interested in that information. You start counting up that type of information and see where it lands, and the Why will practically jump out at you out of the folder. It is so easy! It just leaps right out. But you have to know what you're looking at.

In writing up one eval, an evaluator verbally gave me more valuable data than she had put into the eval. She was quoting reports. All you want to do is quote the steps of your investigation.

The Why has got to be specific. If a Why is insufficiently specific, it just can't be operated.

There's an admin Why, which is the normal one that you're trying to handle. There'll be an admin or tech Why and below that there'll be an ethics Why and above that there'll be a bright idea.

You have a criterion when you've got your evaluation all done, your handling has got to be bright-it's got to be a bright idea, that will actually drive those stats up-and something which can be operated. And if you do an evaluation that cannot be operated at this stage of the game, you're just wasting your time. Look at your resources. What can you do with what you've got? While you improve what you've got. It will all have to be done by a gradient. So the worse off things are the brighter you have to be.

When you do evaluations, you've got to be able to operate the resulting actions. If you write something that can't be operated nothing will happen. That at once tells you whether you have a good evaluation or a bad evaluation.

Do your evaluations in such a way that they are dead on-bang! bang! bang!-­and then, that being the case, they have got to be something that can be operated. And the next thing you know your stats will go up.