How To Organize Keywords Into Ad Groups

Google welcomes its newest AdWords advertisers with a series of emails that comprise an introductory course to paid search advertising.  One of their mini-lessons regards keywords.  Laura from The Google AdWords Team writes:

Then, group your keywords into close-knit themes and create a new ad group for each theme. Put your keywords into these new ad groups. For example, if your campaign is for digital cameras, you can group together mini digital cameras in one ad group and SLR digital cameras in another.

This is not bad for a beginner approach to paid search.  However, if you approach your ad group creation from this standpoint, how will you optimize your spend?  That is: some of the products and/or services that you offer are more valuable than others.  So when Google suggests “close-knit themes”, I’d like to nuance this by encouraging you to knit your themes together, in part, by the margin of the product line in question.

What do you mean, margin?  Is that, like, butter?

Let’s say you are buying keywords for both mini digital cameras and SLR cameras.  And let’s say that you have two different landing pages / destination URLs:

1. A page all about mini digital cameras
2. A page all about SLR cameras

Obviously, you’re going to drive all your keywords regarding mini digicams to the former, and SLRs to the latter.

But what does your traffic do once it arrives on each landing page?  Let’s say 12% of your minidigicam traffic end up buying a minidigicam, and this toy costs them $199, and that nets you (I am making these numbers up) a NPV of $42.

Compare that to your SLR camera traffic.  These puppies have a higher price point, a slightly different target audience, and thus exhibit different funnel throughput numbers.  So, let’s say 5% of your SLR camera traffic ends up buying an SLR, and this toy costs them on average $628, and that nets you NPV $117.

Now, do you want to bid the same amount on “mini digital camera” as you do “digital SLR camera”?

I would encourage you to not necessarily do so.  And despite my suggestion that thinking/marketing does not offer the completion of math homework as a service solution, let’s do some maff:

• 1000 clicks on your minidigicam paid search ad x 12% end up buying = 120 sales
• 120 people buy x each worth $42 = $5,040 NPV
• $5,040 / 1000 clicks = $5.04 per click

So when it comes to buying traffic for minidigicams, if you spend $5.04 on each click to the minidigicam landing page, you’re breaking even.  Spend more?  You lose money.

(This, of course, is a very rough approximation — there are more variables to consider than simply the max bid, particularly if the service and/or product that you offer is not a commodity.  For example, if you market financial services products, the position of your paid search ad can often correlate with customer quality.  But a camera sale, in general, is a camera sale.)

Let’s compare this $5.04 break-even max bid for minidigicams to that for SLRs:

• 1000 clicks on your SLR ad x 5% end up buying = 50 sales
• 50 sales x average net of $117 = $585 NPV
• $585 NPV / 1000 clicks = $0.58 CPC

So even though you’re making more per sale on SLRs, because the sales cycle is more difficult (as represented by the conversion rate of 5% vs. 12%), you actually have to pay considerably less per click in order to just break even.

What does this have to do with “close-knit themes”?

Now, Google’s suggestion for grouping keywords — thematically — is not incorrect.  But I do want to suggest that as you come up with your ad group “themes”, you ensure that one of the dimensions and/or variables of your themes is the value of that click.  And the value of that click is a function of

• the conversion rate of that clicker, once they arrive on your site
• the net present value of the converter (i.e., how much that sale is worth)

Google has a motivation to get you to not think about the value of traffic to you; they have an incentive to get you to forget about how you can monetize that traffic and instead get caught up in a frenzy of bidding against other advertisers who also want the eyeballs of people searching on “mini digital camera”.  So someone out there, if these fake numbers are real, might let the AdWords market — rather than their own margin numbers — dictate how much to spend on “mini digital camera” and end up getting hosed (i.e. spending $17 on a click that they only end up making $5 or so on the backend).

Do not let the price of keywords dictate your maximum spend! (Unless, of course, like Jay Walker’s wallet in the days of priceline’s WebHouse club, you want to bleed cash.)  Let your own internal margins inform what you are willing to spend as your max bid.

(How this may or may not relate to your willingness to abandon your own needs for the needs of others is outside of the scope of this post ;)

Of course, all of this assumes that you:

• Know how much you make on customers who purchase product A vs. product B
• Know how much harder it is to sell product A vs. product B
• Are willing to suspend active marketing of unprofitable products

Does this make sense?  Have I skipped something?  I may have breezed over a few assumptions without realizing it, so please let me know if anything is confusing and/or doesn’t make sense.

And – any other tips / suggestions / ideas that some of you other AdWords jocks might have for folks just getting their feet wet with AdWords?

[ Click here to learn more about thinking/marketing's paid search solutions. ]

Perplexing AdWords screen shot captured in August of 2009 [view]

Tagged as: AdWords, Paid Search, search engine marketing