I am responsible for deep, data- and analytic-centric decision support for our paid search campaign managers. In this role, I can't tell you how many times I spot startling trends in our paid search tracking reports – "Sales are way up!" Or "Traffic is way down!" For one of our clients, the conversion rate rose over 50 percent from the previous month and (more importantly, because of potential seasonality) year over year. "Wow! Conversion rates are soaring!"
Findings like this are then immediately followed by the question, "What did we do?"
Most of the time, of course, there is a direct explanation. It may be the fruit of a carefully designed plan of bid adjustment, ad copy testing, etc. But sometimes the good (or bad) news creates a great deal of head scratching ... "Hmmm. Our bids did not change, and our position in the search results did not change either. Budget? Ad copy? Nope." In other words, "We did nothing."
In those cases, the next steps are to check the data again, then call the client – maybe their Web site went down, or they changed their landing page content without telling us. Then even more troubling questions start to surface – "They didn't change their prices, did they? I saw a new ad for them on TV last night; I wonder if they're launching a new TV campaign?"
You know the story, and the second-guessing that goes on, which inevitably leads us to ask, "Why don't they tell us about that media campaign/Web site adjustment/etc.?" Of course, we know the reason: they're busy and haven't thought about it, or it's not their responsibility/silo.
But as those questions and non-answers swirl around, I'm stuck with the original problem of figuring out what happened. And how can we best do our job – planning and implementing effective paid search campaigns – when we don't know what the client is doing with its other media? The short answer is that we try our best, but recognize that it's not the best situation.
What's In the Mix?
Without knowing the full media mix of TV, radio, newspaper, circulars, DM, etc., we're limited when it comes to optimizing paid campaigns. So, what can be done to improve the situation?
First of all, we need to gather as much information about the client's historical and future media plans as possible. In what media is the client active, and where? How much effort – gross rating points, reach, distribution, etc. – is being put into the campaigns? When are they starting and stopping? What is the message? Is there a new product launch? Is the price changing? Whew. That's a lot of data (and pleading with the client for the information). And did I mention the bewildering number of formats the data comes in? But it's a start.
When we're armed with at least some of this historical data, we can begin to explain the mountains and valleys in those paid tracking reports. We'll be able to see that click-throughs sky-rocketed because of a 2-week TV blitz, or that impressions and clicks increased on branded terms because of a direct mail drop at the beginning of the month. But it gets even better...
As everyone should know, it's important to at least consider the impact of future media plans on your paid marketing campaigns. Should we be increasing our budgets to accommodate higher volume? What new keywords should we buy in anticipation of the product launch? How should our ad copy be adjusted to accommodate the new marketing environment? These are basic questions we ask ourselves, and which should have straightforward answers.
Woe and tribulation for any paid search campaign manager without an interest in those basic answers, because money is sitting on the table (sales for your client and commissions or kudos for your agency). Our account managers, of course, are all over this. Things like holiday and seasonal promotions, keyword expansions, bid adjustments, or campaign fine-tuning are their bread and butter. For special promotions, we oftentimes add in excess of 1,000 keywords, and do daily bid and budget adjustments to accommodate the volatility of the marketplace. It goes without saying that our paid search campaigns are adjusted to a basic business fact – change is the name of the game.
But can we really say those upswings in search activity are proof that other media do, indeed, cross over into the search channel? Was it because we boosted our paid search budget and bids? Or were our efforts really just a side show, and essentially just "tagging along" on a surge in activity that would have happened even without more search resources? These are the hard questions that make someone in my position, as the leader of our analysis and decision support efforts, sweat a little.
Giving Search Its Proper Credit
Maybe that increase in the conversion rate occurred because we raised our bid and position, and not because it "helped boost" the effects of the direct mail campaign. Maybe that upward trend in impressions is the result of better budget allocation, or maybe it's the incremental roll-out of a TV campaign. Which is which? And how can we identify the pivotal role of the paid search campaigns?
Here's where I reach into my toolbox of statistics and predictive modeling. (This is heady stuff, and not meant for the faint hearted.) But here's what I do: Using the historical media mix to build a forecasting model, and given "inputs" or leading indicators of media activity, I identify the anticipated movement in search metrics. With the right data, I can tell the client, "If 2.5 million direct mail pieces drop at the beginning of the month, impressions will increase by 14 percent, the click-through rate by 5 percent, and the conversion rate by 4 percent."
Furthermore – and here's where things get really exciting – I can "play games" with the forecasting model by inputting different bid and budgeting scenarios, and finding the best combination to work in the context of the upcoming media mix. I can tweak the numbers to determine what would lead to the best results, and find out, for example, that if we increased our budget to accommodate the surge, and also increased our bids by 5 percent, we could expect clicks to increase by 11 percent and our cost-per-conversion to drop by 6 percent.
At that point, it's "plug and play" – compare a range of different scenarios, and select the best one, taking into account budget limitations. Now that's what I call smart paid campaign management.
So, when you're designing and implementing your paid campaigns, be sure to remember, "it's all in the mix."
Pat Stroh is VP of analysis & decision support at SEM agency Impaqt. He spearheads analytical initiatives and decision support activities for Impaqt's search marketing campaigns, with a special focus on ROI-optimization.
We report the top search marketing news daily at the Search Engine Watch Blog. You'll find more news from around the Web below.
- Simple tips for tracking campaign performance for offline businesses, Inside AdWords
- Google Maps is spying on my cat, says freaked out BB reader, boing boing
- 21 Reasons Why You Do NOT Need Web Analytics, Web Analytics World
- YellowPages.com to Invest Heavily in New Sales Staff and Locations, ClickZ
- Zango Points Finger at Google in Suit Against PC Tools, ClickZ
- Startup Search: tracking the web startup ecosystem, Niall Kennedy
- It looks like we're being evil, but we're not, Information World Review
- You Can't Fake Real Content, Search Engine Land