Comments on: Bid Management Is Dead

By: George

George — Tue, 15 Mar 2011 18:18:00 +0000

JC, love ya man, but you’re mistaken here. You don’t need to test bids. With copy and landing pages the advertiser is guess what might work well, and with educated guesses should do a decent job, but tests are easy to set up through ACE and can give a clear answer. All good. Testing bids is not what bid management is about. Bid management is about properly measuring the value of traffic coming in and predicting factors that will effect that value (seasonality, time of day, inventory, offers, geography, etc.). Determining the bid is then just simple math: value * fraction of value you’re willing to spend on marketing. The hard part is measuring the value of the traffic accurately, and measuring the effects of other variables to help predict. Even that is fairly straight forward on high traffic terms, the hard part is the torso and tail. Bidding done well involves NO guess work, only scientific measurement. There’s nothing to test. We know this is right because RKG and others have done the enormously difficult math to determine that position has no material impact on traffic value.nnThere is no way to test bids, but thankfully there is also no need to test bids. “All” that’s needed is careful measurement and some sophisticated math to figure out how to predict the quality of traffic on the lower traffic stuff.

By: Jc Zhang

Jc Zhang — Tue, 15 Mar 2011 00:28:00 +0000

I agree. nnThe point is that the potential for conversion optimization in keyword bidding is being given more weight than its worth in the industry. nnWith the amount of variables involved, you can easily get lost in the data and spend too much time trying to figure out the best bid is; whereas optimizing in other areas, such as ad copy testing, give you definite results that you can be confident about. nnThat’s what this is mainly about right? Prioritizing based on the volume/conversion potential of testable events that you can be confident about. nnIf you could perform scientific A/B tests on different bids, you could find out what kinds of conversion/volume differences you have based on differences in max bid. However, the closest thing that does this is Google’s Campaign Experiments, because only the engine can control distribution such that an unbiased test is run – by splitting bids randomly with each impression. Outside of this, there is no way to effectively test two different bids.nnSo two things:n1. Bidding is hard to test scientifically.n2. Therefore, the efficiencies gained from bid optimization is unknown.nnCompare bid testing to the following: n-Ad copy testingn-LP testingn-search-query KW expansion (negative or non negative)nn The difference is that there is a methodology to these tests, the results are clear, and what you learn from these results can be used in the future. With bidding, you are trying to analyze a dynamic auction environment with many variables, known and unknown. Whats the result of focusing on bidding analysis? Volatility. Ambiguity. Confirmation bias. And likely a lot of wasted time.

By: Anonymous

Anonymous — Tue, 15 Mar 2011 00:22:00 +0000

Thanks for doing an expose, and being a thought leader in this space. nnI think the whole “bid management” concept has been a mistake from the beginning. nnI think this issue is one that has been affecting the stock market industry as well – where “quants” as Nasseem Talib, author of “The Black Swan” would say, think they can take everything and put it nicelynnWe’ve never focused on bid optimization for our clients – as that may cause a minor day-to-day improvement, but it won’t make the account improve on a consistent basis for the long term.What will is analyzing all of the other core elements, including time of day, geography, search query match / match type etc.nnYou’ve done a great job at Click Equations on focusing on all of the other elements, especially quality score.nnI did a presentation on this a couple of years ago, where I looked at the different elements of PPC Campaigns, and which ones were taken into account in a bid management tool. The bid management tools are sadly dumb on many of the elements. (They have smartened up a bit recently, but still, they don’t take into account alot of the elements).nnThe elements were by no means comprehensive, but I believe may be a good conversation starter:nhttp://www.slideshare.net/doj464/bid-management-farcennYou’ve got a ton more of the elements that you talk about across your blog posts etc.

By: 10 Steps To Setting Initial Bids

10 Steps To Setting Initial Bids — Mon, 14 Mar 2011 14:56:02 +0000

[...] Craig Danuloff wrote a very interesting, somewhat controversial piece a few weeks back arguing that advanced bidding strategies should take a back seat to core campaign development, keywords, adgroup architecture, ad copy and landing page assignments for most advertisers. We agreed that good bidding can’t save a badly constructed program, but as Bradd Libby correctly observed: you have to start with some bids and doing that well rather than badly is important. [...]

By: George Michie

George Michie — Sun, 13 Feb 2011 15:07:00 +0000

Kudos for stirring the pot, Craig, but I disagree with you here. nnLetu2019s take the case of a hypothetical poorly conceived ad. By ad I mean the whole nine yards: keyword, match type, copy, landing page, negatives, syndication, geo-targeting, everything.nnFor whatever reason for this ad some pieces of this puzzle are not well done, and that results in the traffic drawn in being of lower quality (and lesser value) than it aught to be.nnShould the bid management system react to the actual value of the traffic and bid only what the traffic is actually worth, or should the bid management system calculate a bid based on what the traffic should be worth if the ad was perfectly constructed? Not only is the latter impossible — how would an algorithm gauge that? — even if it could be done it would be unwise to do so in that youu2019d end up paying more than the traffic is currently worth with no guarantees that the problems with the ad will be addressed soon.nnFlushing money down the toilet as we learn is inevitable to a degree, but I’d prefer to have the bid management system spend appropriate amounts given the current quality of the ads and constantly work to improve the ads.nnNo question, smart bidding isnu2019t the only thing, but it is the surest control we have for making paid search cost-effective.

By: SEO Philippines

SEO Philippines — Sat, 12 Feb 2011 03:36:00 +0000

oh we all do. thing is, it’s the same battle field. search engines. whether you are in SEO or PPC. but if somebody there’s a as you say a one correct process for doing things, it’s gonna be great!

By: Richard Fergie

Richard Fergie — Fri, 11 Feb 2011 12:29:00 +0000

“Just like SEO, there’s no blue print for PPC”nnYou are probably correct about this, but I think one of the distinguishing features of PPC Managers vs SEO’s is how much we like to turn things into repeatable processes.nnEven though I agree with you, I am not going to stop looking for the one correct process for doing things

By: Friday PPC Roundup – 16 Links & Bid Management Joins Keywords in the AdWords Obits Edition

Fri, 11 Feb 2011 08:09:16 +0000

[...] two of the smartest PPC blogs in the space. Craig Danuloff from the Click Equations team called bid management dead, which prompted a rebuttal from the RKG blog. The dispute really isn’t as juicy or [...]

By: SEO Philippines

SEO Philippines — Fri, 11 Feb 2011 00:46:00 +0000

Bid management would still be an important and integral part of the search management process. We, the PPC campaign managers, just have different ways in doing it. Many ways to kill a cat you know. I read Mark’s reply as well on RKG, http://www.rimmkaufman.com/rkgblog/2011/02/10/reports-of-bid-managements-death-are-an-exaggeration/, before posting this comment. And I don’t think any of you guys are wrong. Just like SEO, there’s no blue print for PPC as well. What works in certain campaign, may not work in another one.nnSome set an initial structure, landing pages, ad copies, a set of keywords and negatives, then start bidding right away. And they go forward from there. Some though wouldn’t start the bidding process until they have everything set up.nnBut it’s the same end goal.

By: Reports of Bid Management’s Death Are an Exaggeration

Reports of Bid Management’s Death Are an Exaggeration — Thu, 10 Feb 2011 21:39:48 +0000

[...] Danuloff just put out a provocative post on the ClickEquations blog titled Bid Management Is Dead. Is it really? I certainly agree that paid search management goes far beyond setting bids, but I [...]

By: Craig Danuloff

Craig Danuloff — Thu, 10 Feb 2011 21:29:00 +0000

Thanks Chad for sharing your experience. You’re clearly a role model that many others would do well to emulate.

By: Chad Summerhill

Chad Summerhill — Thu, 10 Feb 2011 20:55:00 +0000

My bid-management is deadu2014literally. nnWe have proprietary software, which has been broken for several months. I do miss it and hope for its return (Iu2019m fighting for programming recourses). Whatu2019s not dead is the knowledge and processes that went into building our bid-management. So, I do it manually based off of good data.n nI pull data into Excel using SQL and our data warehouse (web analytics, PPC, revenue data, call center data, etc.), analyze, and bid manually.nnLast year was an awesome PPC year for us without our wonderful bid-management software working consistently.n nHereu2019s what we focused on that led to another successful PPC year:nnu2022tAccount structure (more granular, more appropriate, more visible, more controllable)nu2022tSearch queries not keywords (negative keywords all over the place)nu2022tAd testingnu2022tUnderstanding and using match types appropriately (getting Google to serve the right ad)nu2022tGoogle content network was a big contributornu2022tLanding pagesnu2022tTurning leads into sales (phone sales improvements, email, etc.)nnWould it have been better with automated bid-management running? Yes it would have, but because we have spent so much time on getting our account in order and u201cbid-readyu201d when we do turn it back on in the next couple of weeks we will be optimizing off of a much better foundation.nnSo, I agree with Craig on this one. I hope everyone starts using the term “bid-ready”.n

By: Craig Danuloff

Craig Danuloff — Thu, 10 Feb 2011 20:01:00 +0000

Thanks for your thoughts Chris. Obviously my headline is a massive exageration, which is the new standard for headlines (as I’m sure you’re aware).nnI do believe – based on observances I get to make every day – that the perception that bidding is paid search management is prevalent. And I think it’s wrong. It’s a key element but when done at the expense of other activities it’s an unproductive one. nnYour point, that large accounts or advertisers need more ‘management’ than they can muster so they use automatic bidding because it is automatic, may be true. But many of the accounts and managers we encounter do have the choice of how they spend their time – and I’m being dramatic to suggest they think carefully about how they spend that time. And I’d hope that if the decision makers at the large advertisers who are happy simply bidding away on non-optimized accounts really understood the issues that they’d find a way to assign or apply some different resources to the job. In other words I’m not sure the folks to take the ‘bid only’ job are really acting in their clients best interests. If you’ve got a hammer, the world looks like a nail…nnBroadly speaking it’s a broad and complicated issue. I doubt we disagree much about the fact that most campaigns could stand a lot of improvement that isn’t better bids, that ideally you’d tune the account in various ways to get good solid data before the bid algo was set to do it’s magic, or that great bid technology is a huge advantage if properly applied. Is that right?nnIn reality it’s not an either or – accounts need good ‘management’ and good ‘bidding’. I’m sure bid-only automation is better than an account on auto-pilot, but clearly not what anyone should recommend. (Not that you are, just to make the point).nnThanks again! n

By: Craig Danuloff

Craig Danuloff — Thu, 10 Feb 2011 19:46:00 +0000

Approved

By: Chris Zaharias

Chris Zaharias — Thu, 10 Feb 2011 19:19:00 +0000

I disagree, Craig. nnYour post assumes that infinite (or sufficient) human resources exist to address non-bidding aspects of SEM, and you also imply that people who put bid mgmt high on the list of SEM priorities necessarily spend lots of time on it. Both of those assumptions are wrong. nnWhat’s in shortest supply in the SEM space is people (both client & agency side) who know what they’re doing. Add up all the Click Equations, Efficient Frontiers, Search Agency’s and Trada’s of the world, and you have enough capacity to, IMO, address maybe 20-40% of the SEM market. For the other 60-80%, tools to automate data-driven bidding is essential, and moreover I could just as easily argue that lack of automation in bid management prevents scarce human resources from being applied to the areas you put emphasis on. nnI’ve been out of the SEM space for over a year now, so maybe things have changed, but I doubt that. The largest paid search advertisers – the ones who’ve been able to profitably scale their spend year over year – are *very* focused on bid management, always have been and IMO always will be, for the reasons of resource scarcity I mentioned earlier.