Posted by Michael Conway, Program Manager (Paid Search)
Remember that when trying to predict the future, you’ll most likely not be 100% accurate, so you must set expectations accordingly. The second thing to remember is that complexity does not necessarily equal accuracy. Translation: when presenting your forecast, set expectations with room for error and make the logic behind the forecast simple.
The first thing to do is research the market a as whole. In December, eMarketer put out an executive summary projecting retail e-commerce growth for 2009- bottom line percentage growth expected to be 4.1%. This growth percentage hinges on online shopping growing 6.2%, but online spending per shopper decreasing by 2%. There are other numbers out there and different retailers may see more growth than others. For instance, Wal-Mart will probably exceed the market because discounters tend to perform better in down economies, whereas higher ticket sites may not perform as well. Bottom line is unless you fit into those two categories, all things being equal, clicks will grow 6.2% and average order will decrease 2% in 2009.
Second thing to do is to break 2008 numbers up by quarter and examine each quarter for any additional opportunity for growth within the program. Is there opportunity for keyword growth? What was the new keyword contribution towards growth by quarter and what direction is that trending? (Smaller product offerings probably do not apply here) Was there a quarter where the program was underfunded? Is there a higher threshold of expected efficiency or CPA that would allow for more activity and potential sales growth?
Leaving these questions out of your exercise might be a good way to pad your numbers, but if you are trying to get as much budget as possible, these are legitimate and defendable justifications for asking for more than what the market validates. After deciding what the market and program will do, apply those numbers accordingly and then work towards continuous optimization to exceed expectations.