Aug 112010

The Paid Search Business: Client Retention


For many years, we operated RKG on the "Field of Dreams" principle: 'If you build it [they] will come!' Alan and I loved the idea that, because the channel was measurable and because we thought we could do this as well or better than others in the space at the time, we figured: 1) clients will come find us; and 2) they'll be happy because the numbers will be good. We focused on the software and the service rather than the marketing and sales under the principle that if you provide a valuable service at a fair price you will do well in business. It is said that some people choose a barbecue sauce because they like the taste of the sauce, others buy it because they like the look of the bottle. We've always focused on the sauce, and that focus has served us well. As the industry has matured, we've come to realize that perception is important as well. We were naive to believe that good results by themselves would keep our clients happy. Expectations do not always line up with the range of possible outcomes, and this, combined with the increasing volume of sales noise with pie-in-the-sky promises, makes retention a constant challenge even for us. We're having to put a bit more energy into "the bottle" these days. But for the plethora of agencies who focused on the bottle the day of reckoning seems near. Cracks are starting to show in the veneer because profit margins are under siege for the weaker players in the space. Acquiring new client accounts is expensive. The sales cycle is long, the RFPs are arduous and, for agencies hoping to distinguish themselves through slick marketing materials and salesmanship, the close ratios are sinking. Those who rent tools not only have a hard time distinguishing themselves from every other shop that uses the same or similar rented tools, the SaaS licenses cost real money, taking a significant bite out of the bottom line. Price pressure has also impacted the top line. As competition increases the days of demanding 15% of advertising spend uncapped are gone. Ironically, in our ignorance, RKG set out to be the high quality, high priced agency...only to find to our shock and dismay that we were the only folks in the industry capping fees and we were in fact the lowest priced agency for large accounts. Agencies based in San Fransisco, New York, Boston, LA, etc. have huge overhead and expensive staff. Couple that with the difficulty in retaining quality staff and the fact that there is a steep learning curve for paid search analysts to climb, and the expense picture and service picture start taking shape. However, the real problem for the "bottle" agencies is more fundamental: they can't retain clients. Because the sauce isn't really very good, the promises made to win the business come back to haunt the folks responsible for client services, and that hard won, expensive sale turns unprofitable because the long term value of client accounts isn't there. Very few of the clients who've left us over the years are still with the firm they left us for -- I can think of only one. Over-promising and under-delivering only gets you so far in life. PREDICTION: We're going to see a number of the weaker paid search agencies implode over the next year or two. The first victims will be those who have no bid management software worth mentioning but try to compete in the big leagues. The next wave might be those who've outsourced their core software. I could be wrong here, but I think the agencies that have their own proprietary tools, access to all the data, and the ability to customize their platforms and reporting have the inside track. There will never be universal agreement on the best barbecue sauce, nor will one agency end up being everyone's favorite. We aren't satisfied with our sauce, and never will be; it is and will remain a work in progress. Our bottle is improving but is a LONG way behind the sauce. We don't mean to use this as a platform for braggadocio, but we do see a future where the folks with the watered down sauce in the fancy bottle end up looking for other products to sell.


31 Responses to "The Paid Search Business: Client Retention"
Very interesting, George. Access to all the data, and the ability to customize our platforms and reporting is exactly why we built our own data warehouse web analytics solution. After we had the right data in an accessible format everything changed--for the better. We could now ask any question we wanted and integrate it easily with our other data sources. I wanted to say that if the quality of content on your blog is any indication of the quality of your "sauce" then I believe that you will prosper regardless of what happens in the industry. Especially if you are working on the bottle--there are plenty of people who need a nice bottle. Bottle appreciators may not compare to your sauce appreciators as customers though. You might see a similar trend for bottle-focused software vendors as well.
Chad, thanks so much for your kind words. We need to grab a beer sometime! Seriously, let's chat, I'd love to know more about your business. We don't really understand how an agency can use a third-party tool and really be a contender for best in show. Particularly in this industry, where understanding and making smart use of data is so critical. Seems like entering a baking competition with a Betty Crocker mix cake. That cake might be better than some, but it just can't win.
Bryan Eisenberg says:
I agree that many agencies are going to shatter in the next couple of years, but not because of the software they license or not. That's like saying a BBQ factory will close because of the mixers, bottlers and label machines they use. It's not the tools but how you use them and what recipe you have. If your sauce sucks and you don't account for changing tastes then you deserve to close up shop. Developing great tools can be expensive and it helps seeing how it many companies use these tools and their processes and recipes to leverage as you keep developing these tools. It's kind of like saying you can't be a good web analyst if all you use is Google Analytics. I know plenty of analysts who have helped their client make millions using GA and plenty of analyst who have wasted their salary being a report jockey with a million dollar installation of a "enterprise" Analytics solution.
Scott Clark says:
You wrote..."Agencies based in San Fransisco, New York, Boston, LA, etc. have huge overhead and expensive staff. Couple that with the difficulty in retaining quality staff and the fact that there is a steep learning curve for paid search analysts to climb, and the expense picture and service picture start taking shape." Your statement about coastal overhead got me thinking. Is there an opportunity for specialty SEM firms (or branch offices) to start to flourish in lower cost areas of the country? Naturally, I am thinking of us here in the "golden triangle" of Lexington, Cincinnati and Louisville. My office rent is around 1/2 what it would be in the cities you mention and the quality of life is very high (I moved here from the Bay Area.) Lexington, for example, is one of the most educated cities in the USA and has a stream of tech talent flowing from local universities. It also has a very convenient airport with non-stops to major cities. There is also much more loyalty to a company here than the other cities - mainly because people want to raise families here.
George - Although I deeply respect your knowledge and the insights into the business that your regularly share here, I couldn't disagree more on this one. I'm biased, as you know, but there is one area where we may have some common ground in our thinking. Let me explain. First, software development is a) extremely expensive and b) a very different skill set than providing services. It seem entirely impractical for all but the very largest agencies to even consider it. At least in terms of full platform development - small point tools or utilities do make sense. Second, craftsmen shouldn't need to make their own tools. For the reasons above they usually don't (guitarists or carpenters or whomever - there are exceptions I realize). They should be able to distinguish themselves with strategy and skill - and by should I don't mean the couldn't make tools (beyond the reasons above) but that they shouldn't have to because they can intelligently and skillfully use the tools available to fantastic results. But lastly, this presumes good tools are available. Tools that are accurate, flexible, and well build (logically and structurally). Here is where I start to agree with you. The current generation of tools is in many ways inadequate. Because it's so hard and expensive to development, and due to lack of imagination in many cases, they broadly don't provide the level of capabilities and flexibility they should. I think they can provide dramatic assistance to most practitioners, but they're far from as good as they could be. I can see how you (as I am often) can be super frustrated by them and just want to build your own to fulfill the potential and vision you have in mind. But I don't think that's realistic but for more than a very few out there. On top of needing better tools in the ways I've mentioned, I would agree that tools need to be less ridged. They should be extensible platforms (none are today) so if a craftsman needed a new element, because they had advanced insights into some element like bidding or testing or a metric, they could be modified and added easily. In other words the platforms should not be 'all or nothing' 'our way all the way' resources. They should allow vastly more serious personalization even to the point of programmed alteration or extensibility. My guess is you would appreciate this option too - the ability and need to only code those relatively small parts that you don't like from large and otherwise well serving platforms? I wholeheartedly agree with your initial observations about the agency business. It has huge structural and marketplace problems. As you elequently said in a recent post, in house doesn't work any better. Tools have limitations as we've both said in different ways. A Change is 'Gonna Come - it has to. Only Google should be enjoying the current environment. Your voice in pointing out that most of the emperors and peasants have no close today is helpful and refreshing. It's clear you'll continue to make a huge contribution to finding a better way.
Bryan, I think the tools in paid search are fundamentally different. It isn't simply a reporting interface, it's actively driving the paid search program through a combination of statistical modeling and heuristic rule overlays. Performance proves that all tools are not created equal because the data inputs are different, the stats modeling approach is different (in some cases non-existent, in a few cases really impressive, but each different), and the ability to fine tune the output is different. We're not talking about a bottling machine; the inputs, the algorithm, and the post-modeling controls are critical ingredients in the recipe.
Scott, we're based in Charlottesville, Virginia, and we like the way you think :-)
Craig, I'd have been sorely disappointed if you didn't comment! Thanks for stopping by. I certainly agree that at this stage of the game it is difficult for folks to build their own platform that will compete with the best out there, though it is also true that wonders can be accomplished with enough venture capital. Most folks in the mid-90s would have said that Yahoo and Alta Vista would have the search engine market captured going forward. You may be right. If the platforms available for rent had the algorithmic horsepower, the flexibility to handle all manner of data inputs, and were open for customization, knowledgeable agencies could likely differentiate themselves from the others. But I still believe agencies like RKG, The Search Agency, Efficient Frontier and maybe one or two others have the inside track. Time will tell. As I've said before: I think Trada and platform providers like Click Equations, Marin and Kenshoo have helped raise the bar in the industry and that's been a boon to everyone. It's forcing the pretenders off the stage, improving the woeful performance of others, and helping to convince advertisers that outsourcing is a viable option even if they've been burned by shysters in the past.
George, I am well aware of the needs and utilities of the paid search tools. I also happen to be an advisory board member for Craig's company ClickEquations. And I don't mean to equate PPC tools to analytics in that they are both reporting tools, just that the value is in the insight and creativity that are enabled by the tools doing the dirty work. Good marketing is about persuading people not about a tools algorithms. Believe me, I have seen my share of PPC accounts by some of those other tools/agencies you mentioned that are costing my clients hundreds of thousands of dollars a year because of their flacid attempts at marketing.
No question, Bryan, it has to be a combination. Superb algorithms and tools are wasted by poor application. The flip side is also true: great analysts can only get so far with mediocre tools.
Huh? says:
Does this mean RKG has a bunch of Monkey's pulling levers and pushing buttons of a fancy machine? Or that the "sauce" is only made up of a great tool? SEM is Search Engine Marketing, not Search Engine Machining. Good Tool + Great Marketer = Great outcome Great Tool + Poor Marketer = Fail It seems if I worked at your agency and read this I'd be concered that you do not care at all for the real driver of success: Great Marketers. You don't mention that piece even once, only that the tool matters for results. I'd call that a big Fail.
Marc Poirier says:
Hi George I agree with Craig and Bryan completely (surprise, surprise!) As Craig said, technology companies need to leave room for agencies to inject their special sauce into the platform. Their own campaign management algorithms, their own special metrics, reports, etc. And they're doing it. To spend millions to try to build and maintain connections with not only all search engines, but also all social media outlets, all RTB display vendors, all those API changes, then supporting your clients with technology issues (there's always issues!) - I mean it just makes no sense at all. There are a good number of well known search agencies who USED to have their own technology. They're now using one or several of the platforms that are available. In my opinion, agencies looking for profitability need to steer clear of the path of building their own infrastructure. They should focus on building their methodology and delivering value.
Thanks for your commentary, Marc -- I'm not shocked you came down on that side :-) I also want to point out that I'm not saying we, or EF or TSA necessarily have better tools than CE, Marin, etc, I'm saying they have insanely sharp folks who have the potential and motivation to build superb tools and that the tools are by definition more flexible because they have control.
Huh, Anyone who reads our blog regularly or who has heard me speak at conferences, or has talked to our clients knows that we do NOT subscribe to the black box approach to machine marketing. Both are required.
Huh? says:
I completely understand, but hope you'll keep in mind every post could be the first someone reads. And this one likely gives an impression you may not have intended. More on topic: I think there are two additional pieces of this puzzle: 1) Specialization- creating tools and marketing are specialties better done by separate types of people. At RKG I assume you have those who create the platform, and those who use it. The 1/2 that creates the platform competes against Marin, Kenshoo, etc and may have a leg up right now. The 1/2 that competes on marketing may have an advantage with the better tool. But how long? 2) Improvements- This is still a very young industry, and the tools are young as well. I assume the improvement curve for a SEM tool will be asymptotic, and tools available now are in the exponential growth stage. This assumption is based on the belief there are a limited number (albeit large) of variables that can be collected and hence a limited number of analysis to complete. At some point all tools, in-house or SaaS, will be approaching the asymptote and at this point whether you have your own or someone else’s tool will matter very little. What will matter is who is using the tool. It may be true that in the short-term using an inefficient tool can either deliver lower results, or it can drive up the costs to deliver good results. And this inequality in the marketplace may very well remove some players. But long term, those who survive this period, will have access to tools that reach their apex and the landscape will likely look like other mature industries where the tool competes on price, and the agency competes on talent.
I think you've hit the nail on the head. If you had asked me in 2005, I'd have said: by 2010 paid search technology will essentially be a commodity and the differentiators will be in how the technology is used. I'm quite stunned that here in 2010 the gap is still as large as it is. No doubt it will eventually shrink, but as you say, this remains a young industry and there is still so much to learn. The pace of development for each platform may depend on how hungry the development team is in terms of testing and learning. I think it's going to be a while before everyone with a material paid search program is using top grade tools to manage the program.
George, sure I’d love to chat. Just shoot me an email at chadsummerhill at In response to your statement, “We don’t really understand how an agency can use a third-party tool and really be a contender for best in show.” This happens when a sauce agency finds a sauce software company or rather software companies. Right now, I believe that there are more sauce agencies/people than there is sauce software. There is a race right now going on among the software vendors—they will be forced to build more sauce. Competition is rising and new players (maybe saucy players) enter the game daily. Bryan, in response to, “I agree that many agencies are going to shatter in the next couple of years, but not because of the software they license or not. That’s like saying a BBQ factory will close because of the mixers, bottlers and label machines they use. It’s not the tools but how you use them and what recipe you have.” Software, like better machinery, allows the maker of good sauce to scale with demand. So, given two equally desirable sauces, the one that operates more efficiently will have a higher likelihood of growth/survival. But there will always be the desire for boutique, hand-crafted sauce. Having said all of that, I think I would be much better off if you, George, and Craig managed my paid search campaigns with Google Analytics and Excel. The software hasn’t evolved to be flexible enough or useful enough. But it will. Craig, I’m going to talk about software now. In response to, “…software development is extremely expensive…The current generation of tools is in many ways inadequate… They should be extensible platforms (none are today).” The cost of developing software continues to come down and with the advances in the Lean Startup & Customer Development movements I think that startups are getting smarter about how the spend their money and when they try to scale—focus is shifting to making sure you are solving the right problems and getting to product-market-fit before trying to grow. Now that it takes less money to start your startup, we are going to see lots of new players who are focused on solving the right problems. I agree with you that the current generation of tools in many ways inadequate. I’ve tried to buy software to make my life easier, but most companies are trying to solve problems that aren’t my problems and they want to charge me $10-20k/month to do it. What I’ve observed over the last few years are companies trying to tackle problems like intelligent bid-management, attribution, reporting, analytics/visibility, cross-platform editors, etc. Bid-management and attribution although helpful are used to optimize what should be a good foundation of PPC best practices and mean little when the underlying structures are weak—we end up optimizing sub-par accounts (Craig speaks about this problem quite a bit). What’s the point of robust, intelligent optimization platforms when your keywords aren’t “bid-ready”? Reporting & analytics for increased visibility (or truth) get us further in the evolution of the software, but in so doing have created mountains of work to perform based on the insights. We need software the does the robot work—that takes action on all of these actionable-insights, or at least facilitates action. ClickEquations recent ‘Keyword Zoom’ feature hints at the possibilities of this approach. The customer is taken to the point of action (based on known best practices gained from improved visibility) and then given a button to push to fix the problem. I like where this is headed. The current platforms aren’t extendible. Software vendors want to solve every problem and want their customers to use their platform exclusively—this doesn’t work for me. What if I wanted WordStream’s keyword grouping capabilities, Clickable’s best practice rules engine, ClickEquations Excel Add-In, Marin’s cross-platform editor, etc? The first company to develop a cross-platform editor that can be extended via plug-ins or modules could be at the center of this new evolution of PPC software development. Imagine being the WordPress or Drupal of PPC software (not that it has to be open-source). Having an extendable cross-platform editor (core software) would foster inovation and specialization and help get us past this one-size fits all solution that everyone is trying to develop and speed up the evolution of PPC software. It's an idea anyway.
I think we've reached violent agreement. Better and more flexible tools are needed, and that demands extensibility. Tools should enhance capabilities and not limit or constrict them. We're not there yet. There's a challenge to get there. There's a different challenge between now and then. Until then, people with super-deep understanding like George and Chadd, combined with resources and courage will create custom personal solutions. For that vast majority not nearly as wise as you two, and I mean this most sincerely, from what I see every day they're better off spending time studying the knowledge you two share and getting the basics right - and honestly the tools we and others provide are probably (kinda unfortunately) good enough. And I mean this for most agencies too. I'm pretty sure 98% of them wouldn't know how to use the tools you two (as examples of the very top) have built for yourselves. Makes me think of people snow-plowing with $1500 skis. Just bringing this around to the fact that the knowledge gap is at least as huge as the tools gap. Easier to solve too but somehow despite 1000's of pretty good blog posts by the folks participating on this one I don't see the world's PPC IQ rising very quickly. :-) One day the smartest guys will assemble and tune (but not have to develop) the best tools) and differentiate by their choices and the way they use them. Pretty clear we're all eager for that day, but it's a long ways off. And as Marc adds, there are 5-10 more channels with different individual capabilities and the interplay coming. Lots left to do.
Craig you are too kind. The future is bright for the folks who can put all the pieces together. It is incredibly difficult to try to make a powerful, flexible system user friendly. We've thought about trying to offer our software as a service, but cannot figure out how to wrap it up and make it user-friendly. You're absolutely right, we have to spend months training our analyst to use our system and it's butt-ugly. Let's get a beer at some point! Possibly the first RKG Blog post with not one but two beer meetings suggested!
Tom Demers says:
Wow outstanding comment thread, sorry to be jumping in a bit late. I agree with a lot of the thinking above and wanted to add a couple quick thoughts. I think there are two big opportunities in the PPC software space that are "lower-hanging fruit" than the daunting task of building an end-to-end super-system: 1) Plug-and-play point solutions for experts - As Chad and Craig mentioned, by building more flexible pieces and making them affordable you can address the up-market players and save them development costs/help them improve results. This doesn't offer the revenue opportunity of building something that costs tens of thousands of dollars a month and "locks the advertiser in", but it also doesn't require the maintenance/development/overhead that developing and maintaining a system like that does, and it allows you an entry point into every single advertiser AND agency's workflow - even those with proprietary platforms, which presents you with a nice potential revenue stream. 2) A platform for SMBs and small-to-mid sized agencies - I realize this wasn't the focus of the post but as Craig mentioned: there are far more PPC advertisers and agencies who couldn't begin to consider building out their own tools who would benefit greatly from a "good enough" platform that offered a lower price point. Google itself is experiencing a lot of churn in the tail of it's advertising base, and if you can build something that makes good (but no where near best-in-class) results accessible for a reasonable price you've just opened up a huge market opportunity. The challenge here is really less about integrating with existing systems, etc. and more about balancing power with usability so that you can get the novice-to-intermediate PPC user to profitability without overwhelming them from a usability perspective (no easy task but potentially a lot less expensive from a development/maintenance standpoint). Anyway great stuff thanks to everyone who contributed! Tom
Tom, Thanks for dropping by and adding to a great discussion. George
As someone who's been managing SEM salesteams for the better part of ten years, I'll add that the debate over what is the right SEM model: 1. Agency w/proprietary software 2. Agency w/3rd party tool 3. 3rd party tool used inhouse will be decided much moreso by the salespeople that engage potential clients than by any other factor, including both client preference and industry debate. At eTail East earlier this week it was clear that most advertisers still don't know which model makes the most sense for them, and when they *do* it's usually driven far more by politics than reason. As is clear from the partisan comments in this thread, I think we can agree that our individual vested interests and our passion to win drive the debate. From there it's just a small leap of faith in Ayn Rand objectivism to agree that that who have what it takes to win the game & truly are committed to winning will, in fact, determine who was right. The tip of the spear is Sales, and to the extent you all have the will to win, you'll need to build salesteams with hunters every bit as knowledgeable as you.
Chris, you make a good point here. In fact, I'd argue that there will always be companies that want to outsource marketing services, and there will always be companies that prefer to manage everything in-house. My point is less about the model-preference and more about then end quality of the tools and service. The sales staff will have more success selling quality than selling crap; and the quality providers (software or software+service) will win out in the end because of the higher long term value of the client accounts won. Reducing churn is key to profitability.
Ken Robbins says:
While I enjoyed this article, I am not in full agreement with your conclusions. We too have experienced the difficulty in maintaining and winning new paid search clients as the business matures. The difference in our conclusion is that you see the proprietary-ness of the technology as the differentiator and I don't observe that to be the big issue with clients. Clients are merely tired of having one-trick ponies as their agency. In 2004, a PPC shop was a speciality surgeon. In 2010 the service is demystified and clients either wish to bring it in house or just let their large agency handle it. This is the same path that database marketing, direct mail and other forms of core marketing take. There is no mystery any more in paid search. You're definitely on target that results matter less than they should, but the ones with the "inside track" are the agencies who have a better (bigger) relationship with the marketing executives- and that's often the AOR. Cheers, Ken President Response Mine Interactive Atlanta
Ken, you make an interesting point. I agree that paid search is no longer new and sexy; and indeed, folks are less interested in hearing about it and talking about it. But it isn't a "solved problem" and the range of quality in service providers is still huge. It's fundamentally too important to fall to the AOR folks who have no idea what they're doing, and few firms can build their own tools for doing it well. No doubt, there will always be folks who prefer to do this in-house, there will always be folks who prefer to hire generalists rather than specialists, but there will also always be a place for the specialists, provided the quality of their service justifies the extra hassle of multiple agencies. Thanks for your comments!
Vee Sweeney says:
The poster Scott brought up an excellent point that I would like to touch more on. It is nearly impossible for a smaller business to grow into a larger business in the large areas, such as NY city, L.A., San Francisco etc. I used to live right outside of NYC and to lease office space within a 150 mile radius of the city would run a couple grand for one small room with no A/C. I moved and things are so much cheaper in western PA, Ohio, Indiana, Kentucky, West Virginia etc. There is actually a very successful distance education software company in my town and their rent for a whole three story building is dirt cheap. Companies would be able to better expand and if the bulk of the business is done online, then does it really matter where the office or person is located?
Since I mentioned our in-house web analytics data warehouse earlier in the comments, I thought I would include a link to my new post on why/how we built our own web analytics. It's a high-level overview. Depending on the interest I may write in more detail in the future.
Hi Chad, thanks for the link! If you're willing to share the details I'd be delighted to do a feature at RKG Blog.
Hi George, I would be honored. I'll email you soon and we can talk about the details of the article. Thanks!


Check out what others are saying...
[...] The Paid Search Business: Client Retention, Rimm Kaufman [...]
[...] This post was mentioned on Twitter by Rimm-Kaufman Group, Bradd Libby, Bradd Libby, AllThingsM, Negative Keywords and others. Negative Keywords said: The Paid Search Business: Client Retention: For many years, we operated RKG on the “Field of Dreams” principle: ‘... [...]

Leave A Comment