In the quest for keyword domination the would be PPC marketer often times looks near and far, high and low in search of the most possible keywords for their paid search campaigns. Ceteris paribus, the more keywords you have the better as you’re casting a wider net; and as you run from short tail to long tail, the average CPC (cost per click) or your PPC keywords should go down while the searcher is no less qualified.
Some caveats are in order however as they relate to the following:
- Match type – broader doesn’t necessarily mean better when you’re dealing with variants of high traffic terms. Segregate any terms that may spend wildly into it’s own campaign where you have budgetary and day parting control, and be weary of excessive impressions with a low CTR. This can negatively impact your quality score in Google, as well as signal many unqualified impressions leading Google to view your ad as less relevant, ultimately leading to a higher CPC.
- Profit Margins – pay per click offers it’s own world of analysis and insight fueled largely by ratio analysis. Whether the click through rate (clicks/impressions) or conversion rate (conversions/clicks). A Wall St. analyst works with similar ratios related to metrics found on a public company’s 10k or 10q filings. Never forget that more primitive measures such as ROAS% (Return on Ad Spend) completely fail to account for the basic fact that a product of service isn’t worth it’s gross to a firm. Staff, fixed costs and depreciation, fines, sourcing costs and other costs whiddle down one’s net margin to sometimes 10% or less.
- With this in mind, especially for an ecommerce PPC marketer tracking revenue it’s just as important to group campaigns by products with similar profit margins. Since budgetary control is granted on the campaign level in Google, and finally now in Yahoo (order level in MSN) keep the control spread over the same value plain in case PPC isn’t practical for a product of that nature financial profile. Within that campaign, due your diligence and pick off the more product specific ad groups that are failing to meet certain CPA or cost per conversion goals. If you took a different approach and organized campaigns based on some other consideration it might be impossible to meet certain revenue/cost goals…here’s a new formula.
Return on AdSpend [ROAS%] – adjusted for net margins
ROAS = Revenue/Cost
ROAS Adjusted = Revenue(net margin%)/[cost]
Assume a PPC campaign generated $50,000 for $12,000 cost, traditional ROAS would equal 416%.
Further assume that the net margin on the goods (in this group with similar margins) is 15%, the new math is below
ROAS Adjusted = $50,000(.15)/ [$12000] = 62.5%
So clearly factoring in a profit margin nullifies the apparent value of this campaign. Now this reflects just a moment in time. Not repeat business, not cross selling/cross promotional opportunities, or even the value of not letting another competitor get the client. The sum of this non-quantifiable value must be calculated at your discretion. But clearly arranging your campaigns based on net margin is a smart idea!
If it’s too late to approach organization this way we can still arrive at a worthwhile value through a weighted average.
Assume you have 4 products, a keyword cost of $50,000 and pay per click revenue as follows:
- Product 1 – $12,000 revenue with a 25% net margin
- Product 2 – $9,000 revenue with a 75% net margin
- Product 3 – $25,000 revenue with a 10% net margin
- Product 4 – $2,000 revenue with a 5% net margin
Adjusted ROAS = 12000(.25) + 9000(.75) + 25000(.10) + 2000(.05) = $12,350/$50,000 = 24.7% YIKES :-p
*you can typically find the product by viewing the sku in your web analytics package. In Omniture you’d have to view a products report to view the revenue and than fill in the weights.
So just to reiterate, since budgets are set at the campaign level, it’s easiest to group keywords that generate sales for products with similar margins so far as they’re not koalas and beetles. Each adgroup should contain focused keywords and ad-copy and this won’t negatively impact your quality score.