The following is a summary of some of the more frequently used decision methods and an explanation of why they are flawed:
1. Follow the clicks: This is the most elementary of all decision methods. Under this method advertisers equate clicks with success. If a keyword is generating a lot of clicks then that ad must be working.
Advertisers who use this methodology will eliminate ads with little click activity and will spend more on keywords that generate the most clicks. This method completely ignores what matters most, PROFIT. Clicks don't pay your bills, profits do. If you're not making decisions based on conversions (at the very least), you will be quickly out of the online advertising game.
2. Conversions mean success: All online advertisers want their advertisements to convert to sales. But all conversions aren't created equal. It is very possible that you can get a lot of conversions but still end up losing money. For example if you're spending $20 to get a conversion but you are selling a $20 item, you're losing money. Conversion rates alone do not tell you enough information to know which ads are working and which are not.
3. Products sales mean related ads are working: This assumption seems to make sense. If I am selling a lot of product A, then my ads for product A must be working. But some of the research suggests that over almost half of all purchase result from an unrelated search. Advertisers that employ this form of optimization are turning off ads that are the real drivers of their sales and spending more on ads related to items sold assuming that that relationship really exists.
4. Return on Ad Spend (ROAS): ROAS optimization requires a technology that can report on the revenue generated by each online ad. Using ROAS is a decent optimization method though it is far from perfect. ROAS ignores the fact that most companies don't have a flat profit margin across all products. Therefore, you basically look at the revenue by keyword, multiply by your average profit margin, and subtract out the cost of the advertising. In some cases you will overvalue the performance on the ad if it sold items that were below your average margin. In other cases you will undervalue the ad if it sold items that were above your average profit margin. Finally, ROAS does not allow you to see the relationships between the products sold and the ad that was responsible.
There is only one decision method that is guaranteed to optimize the results of your online marketing campaigns, profit-based optimization This is best of all optimization methodologies; in fact it is the perfect method. Under this method an advertiser matches each sale and the products sold in that sale to a specific online ad. This method allows the advertiser to take into account the unique profit margin that each product has instead of using a flat margin in the ROAS method. By using the specific margin on each item sold an advertiser will never overvalue or undervalue an ad. They will be able to determine exactly how much profit each ad delivered.
When you understand which ads are your most profitable, then you know where to spend the majority of your ad dollars. You also know which ads are not worth investing in any longer because they are not profitable. This method removes any assumptions that the other methods force you to make.