Jan 72009

2008 vs. 2007 Q4 Paid Search Results

With 2008 behind us, we're reviewing results and planning for 2009 with our clients.

Here are some aggregate charts of our clients' results in the fourth quarter of 2008 vs. 2007.

For this analysis, we considered clients who we've served for over a year, who are online retailers (eg omitting lead-gen clients, omitting CPO clients), and who have significant online spends.

Chart one shows the change in PPC-driven sales and cost, 4th quarter 2008 versus 2007.


The plot shows a great deal of variability in year-on-year Q4 sales.

35% of our clients showed positive PPC-driven sales growth in the 4th quarter of 2008 vs. 2007. George Michie points out that some clients showing sales gains did so via significant discounting. 65% experienced a decrease in PPC-driven sales.

On the advertising cost side, 25% of our clients increased their PPC spend 2008 Q4 vs. 2007, while 75% pulled back PPC spend.

The regression line has a slope of 0.59, indicating that on average clients opted to run their campaigns to a lower economic efficiency, instructing us to spend a greater fraction of resulting sales on advertising.

Chart two shows the same data with each point scaled to client revenue. This chart shows that larger clients stayed closer to the origin. Smaller clients were more likely to make larger percentage changes in their advertising strategy, up or down.


We believe it is essential for all online advertisers to segment their online paid search programs by brand vs. non-brand keywords. (More on why and how in a coming post.) We believe that the non-brand term portfolio provides the truest sense of incremental PPC-driven sales.

Charts three and four below present the same information as charts one and two above, but this time only considering non-branded paid search spend and resulting sales.



The regression line is a bit steeper, but is still less than one, indicating again clients choosing to run their campaigns to a lower economic efficiency, spending more to drive each revenue dollar.

Chart five asks this question: "Did competitive (nonbrand) PPC-driven sales fall off more or less than non-PPC sales?" The y-axis presents change in non-brand PPC-driven sales, 2008 Q4 vs. 2007, just as in charts three and four. The x-axis plots total client web sales, excluding sales driven by paid search.


The slope is less than one, indicating that for most clients sales from competitive search were down more than the rest of their site. We believe this is a function of increased competitiveness in PPC, making it harder to stay on the page efficiently. For some clients, heavy use of affiliate coupons this year also pulled orders from PPC.

More about RKG here, more holiday 2008 PPC data here and here.


8 Responses to "2008 vs. 2007 Q4 Paid Search Results "
Wow, you guys and SearchIgnite have Q4 '08 data that couldn't be more opposite. They're saying retail PPC spend grew 12% YoY in Q4, and you're saying 75% of your retail clients spent less YoY. Any thoughts as to why RKG & SI data might be so different? Is it SEM mgmt styles, sample sizes, something else?
Hi Chris, Search Ignite is really just a management tool used by their client retailers, so it's hard to speak of a Search Ignite management style. The Search Ignite tool is a position based bid system, so it's quite likely that retailers using the tool simply got caught up in the bidding wars and the spend increased without sales keeping pace. For our clients it's very much more a function of their marketing decisions. Faced with the choice between spending inefficiently and maintaining efficiency at reduced cost and top line sales, most of our clients chose the latter. Could also relate to differences in the client base. Our clients pick us because they're interested in advertising aggressively but within efficiency limits -- that's kind of our specialty. Search Ignites clients may be more interested in top line than bottom line, they may not be died-in-the-wool direct marketers like the majority of our folks. What did y'all see at Omniture?
Amanda Bird says:
Hi Chris, George, SearchIgnite is used by leading digital agencies and direct clients to manage some of the most sophisticated retail campaigns in the country including JCPenney, Office Depot, Borders and many others. We have an advanced bid management platform which includes predictive modeling, portfolio optimization and rules-based bidding. While our clients are very focused on their ROI from search, they are also predominantly multi-channel marketers who are making paid search budget decisions based on the relative ROI of search marketing in relation to other marketing channels. This could explain some of the difference in the budgeting decisions of our client sets.
Thanks to both of you for the comments, but Amanda, I still feel like your firm's read on search in Q4 needs a bit more explanation to it. I'm inclined to put more faith in RKG's data since they go so deep in explaining what they saw as well as their methodology. That said, samples are samples, and perhaps it's just chance that RKG and SearchIgnite data is so disparate. FWIW, though, I'd love to hear more about how SearchIgnite arrived at those #'s. George - Omniture's not yet aggregated any of its customers' SEM data for publication. I'd like to do that, though, and will let you know if we do.
Hi Amanda, Thanks so much for your comment, I certainly meant no disrespect to the SearchIgnite tool! I have had the tour and while I'd make the case that "leading" agencies, by definition, can't rely on someone else for the central piece of their service offering, we've actually recommended your tool to agencies who've asked to rent ours (which we don't do). I think your analysis may be right on. The multichannel folks who are really brick and mortar first and foremost have store support in mind and other big ticket media expenses that play a role in how they manage ROI targets. Our Brick and Mortar chain retailers were also more likely to see YOY "growth" at the expense of observed online efficiency. Could it also be that SearchIgnite considers small agencies as "customers", so that the growth of those agencies' businesses from adding more clients shows up as YOY growth of SearchIgnite customers? Thanks again, Amanda. Chris if you'd share that would be great. May 2009 be better than expected for all of our clients!
Amanda Bird says:
Chris - SearchIgnite’s data methodology is the same that we have used for all of our past reports, in that we only show data on a same-client basis. Our latest report focused on our retail clients. This means that the Y/Y data reported represents all retail clients that were with SearchIgnite for the full Q4 2007 and the full Q4 2008. George – We break down the data by customer account regardless if they are direct clients or agencies therefore if an agency added a client within 2008 that client would not be reflected in the report. Thanks for the recommendations! We obviously take a different view in that we feel that most agencies cannot afford not to trust their technology needs to a leader in space that focuses on nothing but creating the best solutions for tracking, optimization, landing page and creative testing, cross-channel attribution, etc. For most agencies, this is simply not a core competency and they are better served focusing on the strategic aspects of campaign management not the underlying technology required to implement those strategies. Rimm-Kaufman’s technology platform obviously serves your clients very well and congrats on your success.
Well said, Amanda. Building the technology from scratch in 2009 would be a heck of a steep hill to climb for a young agency. I think you have to be the best technologically and have complete control of the platform to be a leader, but you don't have to lead to be good. Strangely, a call from Marin Software just came into my VM :-) In point of fact, we recommend that small agencies who don't have the knowledge, skills and time to invest in building a top notch platform look at SearchIgnite, Marin Software and Omniture as solution providers.


Check out what others are saying...
[...] Rimm-Kaufman, who is incredibly data driven, posted an aggregate data analysis of his client’s Q4 2008 versus 2007 search ad spend. Let me bullet [...]

Leave A Comment