In some lines samples would be wasted on children, and they are most apt to get them. Then say in your coupon "adults only." Children will not present such coupons, and they will rarely mail them.

But one must be careful about publishing coupons good for a full-size package at any store. Some people, and even dealers, may buy up many papers. We do not announce the date of such offers. And we insert them in Sunday papers, not so easily bought up.

But we do not advocate samples given out promiscuously. Samples distributed to homes, like waifs on the doorsteps, probably never pay. Many of them never reach the housewife. When they do, there is no predilection for them. The product is cheapened. It is not introduced in a favorable way.

So with demonstrations in stores. There is always a way to get the same results at a fraction of the cost.

Many advertisers do not understand this. They supply thousands of samples to dealers to be handed out as they will. Could a trace be placed on the cost of returns, the advertiser would be stunned.

Give samples to interested people only. Give them only to people who exhibit that interest by some effort. Give them only to people to whom you have told your story. First create an atmosphere of respect, a desire, an expectation. When people are in that mood, your sample will usually confirm the qualities you claim.

Here again comes the advantage of figuring cost per customer. That is the only way to gauge advertising. Samples sometimes seem to double advertising cost. They often cost more than the advertising. Yet, rightly used, they almost invariably form the cheapest way to get customers. And that is what you want.

The arguments against samples are usually biased. They may come from advertising agents who like to see all the advertising money spent in print. Answer such arguments by tests. Try some towns with them, some without them. Where samples are effectively employed, we rarely find a line where they do not lessen the cost per customer.