Pay-per-click search advertising (with Google AdWords, Yahoo Search Marketing, MSN Ad Center, or Ask.com) has been one of the fastest growing marketing channels. Yet, since paid search has only been around for a few years most search marketers are frankly still winging it when it comes to optimizing how much they bid to pay on a clickthrough from each keyword. Many companies outsource their search marketing to agencies in the hopes that those will do a better job. Yet, many agencies frankly also just wing it. Take this conversation:
=========================================================
Search Marketer: So, Akin, should I bid more on keyword "Color copier" or "Photocopier X2000"?
Akin: Let me pull up my NetInsight web analytics solution and find out for you which keyword performs better so that you can maximize your return on ad spending. Let’s see which keyword converts better on your site:
Color copier keyword: 200 visitors completed a purchase
X2000 keyword: 100 visitors completed a purchase.
Search Marketer: Oh, thank you. So I will bid more on "Color copier".
Akin: But wait, NetInsight also reports conversion rates, i.e. the ratio of total clickthroughs to the subset of those who converted by completing a purchase. I see that:
Color copier: 10,000 clickthroughs, 200 purchases, i.e. 2% conversion rate
X2000: 200 clickthroughs, 100 purchases, i.e. 50% conversion rate.
Search Marketer: Oh, I see. Since my ad rank for X2000 is lower I should bid more on this keyword then because the visitors that click through are much more likely to convert? They already know they want an X2000 after all.
Akin: Yes, but wait. We can also take a look at your average order value from visitors that come to your site searching for each of these keywords:
Color copier: Average order value $500
X2000: Average order value $300
Search Marketer: Oh, boy, so I should bid more on "Color copier" because those visitors buy more from me?
Akin: Yes, but wait. We can also see the pay-per-click costs in NetInsight so that we can gauge how much you are paying vs. selling.
Color copier: Average cost of clickthrough: $8
X2000: Average cost of clickthrough: only 50c
In other words, cash flow on color copier has been:
Color copier bid costs: 10,000 x $8 = $80,000
Color copier revenue: 200 x $500 = $100,000
--------
Color copier contribution: $20,000
Likewise the cash flow on the X2000 keyword has been:
X2000 bid costs: 1,000 x 50c = $500
X2000 revenue: 100 x $300 = $30,000
--------
X2000 contribution: $29,500
Search Marketer: Oh, boy. So I should bid more on "X2000" because I get much more contribution?
Akin: Yes, but wait. I see that the X2000 product itself costs you $289. (This is a loss leader product.) So the return calculation for X2000 now is:
X2000 bid costs: $500
X2000 revenue: $30,000
X2000 product costs: 100 x $289 = $28,900
--------
X2000 return: $600
Likewise the color copier product commands a higher profit margin as it only costs you $75 (you bought them in bulk from a foreclosure). So the return calculation for color copier now is:
Color copier bid costs: $80,000
Color copier revenue: $100,000
Color copier product costs: 200 x $75 = $15,000
------
Color copier return: $5,000
Search Marketer: Fascinating. So I should bid more on "Color Copier" after all because I get much more return?
Akin: Yes, but wait. From reading the following book I know that the key metric for optimizing marketing is not return but ROI, i.e. returns divided by marketing investments and expressed as a percentage.
X2000 return: $600
X2000 marketing investment: $500
ROI: 120%
Color copier return: $5,000
Color copier marketing investment: $80,000
ROI: 6.25%
Search Marketer: Oh wow, so if I put my $80,000 into the X2000 marketing campaign instead of the Color copier marketing campaign I might get 120%, i.e. $96,000 in returns instead of just $5,000 returns from the color copier campaign today?
Akin: Yes, but hold your horses, dude! First of all, that depends on how many people search for X2000 and are looking to buy one. It is not immediately clear whether you can in fact spend $80,000 on that keyword in the same time period that we are looking at here. A quick reference to keyword research tools can get you an answer.
Most likely also your conversion rate may go down as your traffic goes up. There is no way other than testing to find out what the new conversion rate may be.
Finally, the book mentioned above also cautions of a common mistake in ROI calculations. Namely, to neglect incremental life time purchases of visitors acquired through the campaign. I see in NetInsight that:
X2000 buyers hardly ever make a repeat purchase. Apparently they hate the product.
Color copier buyers on the other hand keep coming back for those expensive color toners (on which you hold the worldwide monopoly). Each of these customers makes incremental purchases that bring you an average return of $500 over the course of the year. This changes our ROI calculation to:
Color copier initial return: $5,000
Color copier incremental return: 200 x $500 = $100,000
Color copier marketing investment: $80,000
ROI: 131%
Search Marketer: Wholly molly. So after considering the incremental life time returns from color copier customers it looks as if I should keep my marketing budget on the color copier keyword after all.
=========================================================
Now, 99% of today’s search marketers seem to optimize their search bids based on conversion rates and initial revenues. Considerations for actual returns, ROI %, and incremental returns are left out. Moreover, offline purchases following online research are neglected too. The search marketer that calculates all this more correctly can figure out what the real price is that she can afford on each keyword. And she will be the one who wins the bidding war.