One of the most important questions advertisers can ask is "where should I spend my next dollar?" (or next million dollars, if you're fortunate). Often, we think the answer is easy. We take a look at our most efficient vehicle, and that is where we drop the next dollar. Then we keep doing it until that vehicle is no longer the most efficient, or we run out of money.
I used to work in yellow pages (yes, I admit it). In the early '90s, publishers really grabbed on to the ROI (define) metric, showing advertisers who had a half page ad that if they bought a full page ad, they would get more sales and still have a really good ROI, sometimes even better than their target.
I challenged one of the consultants hired by the publishers to measure the incremental ROI, not just the net. I wanted him to show me how much the difference in return would be between a half page ad and full page ad, and compare that to the added costs. He told me that he wouldn't do it and then wouldn't respond to my next question, which was, "Is that because the incremental ROI is not so good?"
While publishers might not like the question, advertisers and agencies need to ask it all the time. What is the point of diminishing returns, and how far can it diminish before the vehicle is no longer attractive for my next dollar? While this question is important year-round, it is particularly poignant during the holidays. You no doubt have done your planning and have the plans set in motion. But, you need to question your results, especially in SEM.
Calculating Incremental ROI
Sticking to a cost per acquisition example, the discussion in planning usually goes something like this. I have several groups of keywords that generate customers, as such:
- Keyword Group A at $25 per sale, generating about 10,000 orders a month.
- Keyword Group B at $35 per sale, generating about 2,000 orders a month.
- Keyword Group C at $45 per sale, generating nothing because it's dark.
- Oh, and by the way, I have a display program that generates 1,000 customers a month at $75.
With a $320,000 monthly budget, I have Group A tapped out and Group B about 66 percent funded. Group C is out of luck, because the client wants to see their name in display ads, which sucks up the rest of the budget.
Along come the holidays, and I now have another $30,000 tacked onto my monthly budget. Where do I put it? Most plans would drop it into Group B, realizing that it will likely drive up the average as they start paying premiums to increase position. So, the number might look something like this:
- KW Group A at $25, generating about 10,000 orders a month.
- KW Group B at $39, generating about 2,564 orders a month.
- KW Group C at $45, generating nothing because it's dark (still).
- Display program at $75, generates 1,000 orders a month.
On the surface, the $39 cost per order for Group B appears pretty good. However, strip out what you would have received in your base spend, and you wind up paying $53 per incremental order. Instead of 564 new orders, you could have received over 665 new orders if you had been looking at the incremental cost of each unit instead of the net cost per unit. How? By capping Group B, and starting up Group C.
The next step is getting to the point where you already know where you'll spend the next dollar based on the incremental ROI, or cost per sale. This requires a long term view, but not necessarily a lot of budget. You just need to keep in touch with the SEM ecosystem.
Like most ecosystems, SEM is actually a sub-system. In this case, sub to advertising. This is important to recognize as it forces us to maintain a sense of perspective. It is possible that the next budget increase should not even happen in SEM. There may be other media or marketing efforts that are more meaningful. For the sake of this discussion however, we can limit the scope to online media.
Spend Based on Results, Not Budgets
In the past, I have been very vocal about the need to eliminate a budget-based approach to SEM. To keep an eye on the ball, everything should tie back to the most important business metric. Usually, this is the sale. If you know that every customer you get is worth $100 and you want to make $60 on each customer, why would you forgo any opportunity that brought you a sale for $40 or less?
Budget-based approaches force you to do just that. But they have an even more insidious affect. Instead of thinking about maximizing what you make on each sale, budgets drive us to take the easy route. So long as I can keep my average below $40, I won't really see the fact that the incremental cost is actually much higher.
As in the example previously outline, it is really easy to pile money into the "$35 per sale" opportunity, even as the average increases, because the average stays under $40 (or whatever your target may be). But the incremental is much higher.
So, how do you know when to add money to existing opportunities instead of diving into new ones? This requires an ongoing baseline matrix program and some flexibility. The matrix should consist of vertical media (SEM, display, email, etc.), their associated tactics, the potential volume and the target metric spread (i.e. $20 cost per sale up to $50 cost per sale).
The flexibility comes in being able to dabble in each tactic throughout the year. Again, this is much easier when you move away from the budget-based programs and move to a cost per acquisition program. I delved into this a while back on my Think About Search blog.
Throughout the year, track everything you do in the matrix. If your target is a $40 cost per sale, and something comes in above, say $60, the impulse is to just drop the tactic and conclude that it "didn't work." When time comes to add to the budget, these tactics may appear out of bounds.
But, when compared to the incremental cost of adding to existing tactics, they may in fact be cheaper. By committing these "failures" to the matrix rather than your subjective memory, you get an objective assessment of the relative value.
Test Your Assumptions
Now we come to the flexibility. If you are truly stuck to a budget-based approach, strongly push for a test budget. Something that allows you to dabble in the things that you know will be slightly too expensive. Then, based on your product and seasonality (like holidays or back to school), you should 'play' with the money throughout the year, keeping track of the cost per sale range, volume and volatility. Commit this to the matrix.
Finally, play with what you know works. Just as we overestimate the value of a tactic when it comes to additional budgets, we can become too complacent with existing tactics. You can test things like drawing down the spend, dropping positioning, or removing some keywords that may be assists and see what happens.
We know the likely effect is a loss in volume. But, what if the lost sales were costing you twice as much as your target cost? Are there other things that you thought were failures that suddenly become more attractive with the newfound 'savings?' Even if you did this before, schedule times to test old assumptions in addition to new tactics.
Optimization is not just about SEM. It is about the entire advertising program. Only by understanding the relative and dynamic value of individual tactics within each strategy can you have a truly optimized program. Work to uncover the truth, flattering or otherwise, and you will not only optimize the SEM program, but you will maximize your value to your client (or your own company.)
Steve Haar is the senior director of media at Leapfrog Online, where he oversees search marketing and other vertical efforts. Haar sits on the IAB Search Committee, SEMPO Global Search Committee, and the steering committee for the Click Quality Council. He combines his past experience in offline branding and mass media with his passion for online marketing in his blog, Think About Search.
We report the top search marketing news daily at the Search Engine Watch Blog. You'll find more news from around the Web below.
- Quantcast Brings on CMO, Expands Outreach to Publishers, Advertisers, ClickZ
- Engagement: The Definition Debate, ClickZ
- Death, Social Media, and Remembrance, ClickZ
- Google is Changing the Definition of a "Click" on AdSense Ads, Search Engine Roundtable
- Microsoft launches Live Search Webmaster Portal beta, LiveSide
- The Web in Google's Mobile Browser, Google Operating System
- Face Off: Facial Recognition Search Engines, Alt Search Engines
- Past Time For MSN To Pony Up To The Real Truth About Referrer Spam, SEO Scoop
- Yahoo on Search Advertising and Behavioral Targeting, SEO by the Sea
- The Role of B2B SEO in the Buying Cycle, B2B SEO
- Voters Are Searching But Campaigns Aren't Advertising: Analyzing PPC Results for 50,000+ Google & Yahoo Political Searches, Rimm Kaufman Group