SEM: Pricing Revolution
Compensation models for SEM firms are changing and must continue to change. Over the next two or three years many of the dinosaur agencies will fall because they cannot adapt.
Changes must come in the pricing models and in the overall compensation amounts, and these changes will revolutionize the industry…for the better!
PRICING MODELS: We’ve long argued that pricing models impact agency behavior. Let’s take a look at the two most common models, and discuss a more rational approach to compensation.
Revenue Sharing
Contrary to popular belief, revenue sharing arrangements do not serve the advertiser’s interests for a number of reasons.
- Too much money for too little work: In paid search, particularly for advertisers who spend a great deal of money in offline advertising, often the lion’s share of the revenue tracked to paid search comes through navigational brand/trademark search. The first five minutes of work for that advertiser — building the brand campaigns — might get the agency 60 – 70% of their pay check. This creates great incentive to grab little more than the lowest hanging fruit and move on to the next client. We often see the evidence of this in data from rev-share rewarded agencies.
- Too many battles over tracking: Who’s numbers prevail in cases of dispute, and how much time you want to spend arguing over the counts are salient issues, here. As multichannel attribution mechanisms evolve this will become a bigger and bigger problem.
- Agencies focus on the highest rev-share percentages: Like commissioned sales, and many partnership referral deals, whoever offers the biggest reward gets the attention. For agencies with clients who are competitors of each other, this can lead to favoritism or pressure to pay larger commissions that the advertiser neither can nor should afford.
- Lost control over marketing objectives: Advertisers go through periods of pushing for top line growth, vs bottom-line profits. Sometimes this is done to qualify for specific manufacturer incentives, sometimes to meet specific goals, whatever the reason: rev-share leaves the decisions regarding aggressiveness to the agency. Changing their compensation levels doesn’t necessarily lead them to be more aggressive, they may decide just to enjoy the extra proceeds.
Markup on Advertising Spend
This is a much better model. The agency has to work to spend money cost effectively. Brand ads spend very little, so the big rewards require doing the hard work of building and managing the competitive search program.
But there are problems here too.
- Incentive to spend: The agency has incentive to spend, perhaps far beyond the point of diminishing marginal returns.
- Work doesn’t scale with spend, necessarily: As a rule, advertisers who spend more demand more of their agencies. That makes sense. But there is a disconnect. Whether a Keyword generates 100 clicks in a week, or 1000, the agency doesn’t necessarily have to do more work to manage it; certainly not 10 times as much. Similarly, whether the average CPC is $0.50 or $5.00, the work seems pretty much the same.
The origins of the markup model lie in traditional agency relationships. The idea being that a block-buster advertising campaign will end up producing huge rewards for the agency that came up with the brilliant concept. Successful campaigns have legs, produce spin-off ads, more advertising and reward creative genius. Mediocre campaigns don’t get very far, don’t spend much money and don’t produce big rewards for the advertiser or their agency.
Here’s the thing: there’s not that much creative genius involved in paid search. There is engineering and smart statistics, analytic excellence and execution…but paid search programs are primarily large or small because of market forces, not because of the agency. The cost of managing a hugely successful program is not exponentially larger than the cost of managing a modest program, and both require similar levels of expertise and advanced tools. Unlike traditional media, the basic size of the program (large, medium, small) isn’t a reflection of the quality or quantity of a competent agency’s work. {A well-managed program will, of course be larger than a poorly managed one, the point is that Amazon’s program will be huge regardless of how well it’s managed}
Ad markup models are often adjusted to rationalize for these effects: by creating a tiered percentage structure, or, like us, creating a maximum fee cap. Neither of these mechanisms is ideal.
PRICE RATIONALIZATION IS COMING:
Ultimately, the agencies compensation must gravitate towards the cost of providing the service. An advertiser’s fee should cover the cost of the analyst time, a share of the cost of the ongoing software development, and some markup over those to cover operating expenses and return on owner investment. Good agencies require and deserve to make a reasonable profit, but the days of agencies making $100K per month or more on large accounts are coming to an end.
Google sees its primary avenue for growth in contextual advertising: be that display advertising, video advertising, rich media or text ads. Growth rates in search advertising have slowed precipitously and at this point are mostly tied to growth in the user base.
However, Google sees some opportunity to grow search revenues centered around two facts:
- They see agency fees coming out of their pocket: if advertisers pay 10% of their advertising $ to an agency, that’s money that could go to the media fees instead. Top agencies will rightly point out that advertisers spend more than they would without our services because we manage the programs to peak efficiency. Google would likely agree with this point, but rightly contends:
- Most agencies manage search poorly: Not only do they shave off a percentage of the advertiser’s money for themselves, they also do a lousy job of managing search programs, such that the advertisers could spend much more within their efficiency requirements under proper management
- Small programs are underserved: The cost of managing paid search well is fairly high, and prohibitively so for small programs. Better self-management tools will allow small niche online advertisers, and more importantly, local brick and mortar businesses to advertise cost effectively without need for a ton of management expense.
Google intends to serve this third group, and to dis-intermediate poor agencies by providing increasingly powerful do-it-yourself tools. We maintain that no black box algorithm will ever do the job as well as a smart analyst with terrific tools and a keen interest in the success of their own account. That said, Google’s tools will very likely run many weak agencies out of business, and as Google’s tools improve agencies charging outrageous fees may find it harder and harder to justify their existence.
For agencies that have grown fat collecting huge fees for mediocre work the day of reckoning is near.
We say: Bring it on!
Interesting post as usual George. Seems to me that the one aspect not mentioned is the trend in-house. For the reasons you site in favor of rationalization, many (or perhaps most) advertisers find that poor quality and unreasonable fees (or inadequate attention) are pushing them to abandon agencies.
I think the great agencies will survive – RKG included – to serve those serious advertisers who both need and are willing to pay for really differentiated abilities. Totally agree that fee structures aren’t right today and think it’s great for you to start a discussion to find the right solution.
But the many agencies whose only service was arbitraging a little PPC knowledge will fall by the way. There are a lot more experienced people to hire in-house these days, and better tools from the engines and us third parties. This isn’t a panacea of course – it’s just as easy to be bad at PPC in-house as it was at an agency! Advertisers will have to pay for value one way or the other.
Thanks for your kind words, Craig.
I agree, there will always be those who prefer to manage programs in-house and those who prefer to outsource to experts.
There are more and more “experienced” folks in the space to be hired, but many times those impressive resumes are held by folks who’ve managed search badly for years for an agency or someone else’s in-house program. Quality experienced talent is still awfully hard to find.
As you point out, doing search well requires both the tools and the expertise, and the outside providers who have only one or neither of those ingredients may have a hard time making money in the future.
That’s a good thing and evidence of a maturing industry.
I completely object to the assumption made in the last comment that to get “expert” attention you need to ourtsource.
Outsourcing provides limited competetive advantage, and there is no inherrent link between “expertise” and an agency environment. Indeed, agencies provide extremely limited expertise when it comes to market places that in-house teams are far closer to.
The priority of an agency is always going to be personal profit, in-house staff work for the overall benefit of a company and it’s online marketing program.
George,
This is a great article.
I think that Google is certainly attempting to better directly serve niche advertisers, but I don’t think that their intention is to squeeze out mediocre agencies – nor will pressure on mediocre agencies be an unintended consequence of this. Google’s revenue curve steepens with each additional advertiser, so many poorly-run, large-budget campaigns in a crowded niche are going to make Google more money. They want as many advertisers in the game as possible.
“Local” and/or “bricks-and-mortar” businesses that don’t directly sell on the web are, by definition, going to be less efficient on a per-conversion basis than their counterparts that do sell on the web. These businesses that run their own SEM campaigns are going to feel the squeeze first, as they are going to have a harder time with conversion attribution.
Tom
Hi Steve,
Didn’t mean to imply that experts could only be found at agencies, far from it. And, indeed the thrust of the article is that many agencies have no experts in the building.
But just as it is the case that some companies have real in-house experts, it is also the case that some agencies put their client’s interests before their own. We are likely a minority, but…
We follow the belief that by protecting our client’s interests first, last and always we end up with long lasting relationships and raving fans who recommend us to others. It ends up being profitable for us, but it is “success by indirection.” We frequently push back when our clients ask us to do things that we know will waste their money and fail to achieve the objectives they seek. We lose the short term windfall but keep the client for the long term.
Arguing that agencies only care about their paycheck is no different than saying employees only care about their paychecks. True of some, but not true of the good ones.
Tom, thanks for your comments.
The engines certainly benefit in the short run from companies managing search badly and flushing money down the drain. I do think in the long run they’re better served by well-managed programs. More on that theme next week!
George
Hi George,
Thanks for the clarity and focus on an often difficult topic.
I agree that the percentage of spend model can exhibit a conflict of interest, as there is an incentive for the agency to ‘get away spend’ which may be inefficient for the client.
I think revenue sharing is an improvement on the percentage of spend model, but as you point out, the natural skew of brand can make an under-performing account look otherwise. Agencies should be rewarded on their ability to add value, but where is the value being added from harvesting a client’s brand equity? A clear separation or benchmark needs to be set for brand vs. non-brand for a revenue-sharing model to work.
You’ve overlooked one other model which I found can be especially effective: the set monthly management fee. With a set monthly fee, there is no incentive to ‘get away’ inefficient spend, the agency can forecast and budget revenue accurately and fees don’t increase linearly (and unfairly – as you point out, 100% extra spend is not 100% extra work).
However, with a set monthly management fee, there is no monetary incentive for the agency to perform (other than to keep the client happy and continue the contract) and it requires a trusting relationship between client and agency to avoid ‘setting and forgetting’.
I don’t think there is a ‘best’ model for SEM – guess it depends on what works best for each agency / client.
Cheers,
Alan
These facts are really commendable…
They see agency fees coming out of their pocket
Most agencies manage search poorly
Small programs are underserved:
Thank you for your comments Dhiraj and Alan.
Alan, our capped model effectively works out to be a flat fee for our larger clients. While there is no short-term monetary incentive to work harder for these folks, we do nevertheless. The reward is maintaining long client relationships which means profitable growth for the agency with no need to burn lots of money on marketing.
Kind of like the interview with Ken Cassar of Nielsen: if you constantly have to “buy” new clients to stay in business, it’s hard to make a profit. If instead, you develop a loyal customer base that refers more business your way, profitable growth takes care of itself.
I do agree with you though, that any reasonable model can be made to work, the more important notion is that the fees should make sense for the work performed.
George – What you neglect to consider about brand bidding is that agencies can drive more sales and create efficiency on existing brand bidding. Optimizing ad copy, landing pages, sales funnels, capturing emails, and promoting special offers designed to move certain products/services/promotions can all greatly increase a client’s return on brand bidding. Sure, brand bidding is low-hanging fruit for agencies, but a good agency will take that cheap, targeted traffic, and improve profitability for the client there, as well.
Hi Josh,
We don’t disagree that there is some value to bidding on brand terms: http://www.rimmkaufman.com/rkgblog/2007/08/23/brand-ppc-a-waste-of-money/
We do think most of this is just cannibalizing sales from natural search and affiliates, but the cost of the advertising is small and the time required to manage is too. Indeed, that’s the point. The relatively small amount of time required to manage that piece shouldn’t generate the lion’s share of an agency’s revenue.
Thanks George for sharing.
When reading through this, the only thing I could think about was Education.
For Web Analytics, we usually have vendor certifications such as the Google Analytics IQ test. And of what I see, Google Adwords also has there certification.
I found an interesting comparison here: http://www.seoservicesgroup.com/blog/2009/03/google-adwords-google-analytics/13070/
I can’t help but wonder if there is a lot of overlap and what that percentage might be of those who are both Analytics & Adwords certified. I’d also be interested by a lot of other variables ;-)
What do you think of the certifications?
Thanks again for your post, interesting read!
Aurélie
Aurelie, thanks for your comment.
To date the AdWords Professional Certification has been a joke. Mostly about familiarity with the UI and rules, nothing about whether the person understands paid search marketing. They’re planning to put a better test in place soon, which they promise will be “hard.”
The issue previously wasn’t that it was too easy, just that you could memorize tons of junk to pass the test without having any clue how to manage a paid search program. Maybe the new test will be better, but the reality is, strange as it is to say, most people at Google don’t know how paid search works from the advertiser’s perspective. There are some folks who’ve started looking at conversion data to learn how this stuff works, but until recently they weren’t really interested in what makes a program large AND efficient.
The test we make our analysts take before they take on a client account is a pretty good model, but we’re not quite ready to share that :-)
“strange as it is to say, most people at Google don’t know how paid search works from the advertiser’s perspective. There are some folks who’ve started…”
I recall an HBR article about MS’s new CFO who started in 2001. Until then there had not been a unified accounting/ reporting system. That blew my mind.
It is hard for us to imagine how a business could become so enormous, so ubiquitous without being managed efficiently or properly.
Same thing here – we’re contending that Google really doesn’t understand the needs of their primary revenue source. Incredible!
Thanks Derek,
In Google’s defense, there’s no inherent reason a publisher should know much about successful advertising techniques. TV advertisers don’t ask NBC for advice on how much to advertise, when, and what the content should look like, nor to catalogers ask their printers and list brokers how many catalogs they should mail, page counts, layouts, etc. Indeed, it’s kind of odd that folks look to the fox for advice on hen house security, and even more odd that the fox sometimes gives reasonable advice.
Hi George,
Thanks for posting. I was agreeing with all of it…until your response to Derek.
Media companies (publishers) *must* know about successful advertising techniques/practices. How can they execute successful efforts without knowing what works?
The fox/hen house security characterization is as true for agencies as it is for publishers, though your limit on earnings is a reasonable safeguard against over “profiting”.
For publishers, in these days of measurable performance, the safeguards are in place too. If the media we recommend doesn’t work, there’s no renewal, as you point out regarding agencies above.
Lets continue this on stage at SMX Advanced!
Hi Chris,
Thanks for chiming in! I may take you up on that panel!
I don’t mean to bash the Engines. Indeed, our agency reps get it as well as they can, but they haven’t historically gotten to see the conversion data that drives the decision-making, haven’t actually built and run a paid search program and few have been ad buyers themselves. As such, their individual experience is pretty limited.
They have gotten better over the years. We haven’t had anyone tell us that “position 1 works really well!” for some time — though some of the satellite office staffers will occasionally plant that bug in one of our clients’ ears helpfully.
Collectively they’re getting it. Hal Varian gets it in a big way.
Hi George,
I’ve just written an article looking at the economics behind the markup model.
http://www.alanmitchell.com.au/techniques/economics-of-ppc-pricing-why-the-markup-model-is-flawed/
Would love to hear you thoughts.
Cheers,
Alan