Feb 102011

Reports of Bid Management's Death Are an Exaggeration

Craig 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 respectfully take issue with some of the finer points Craig makes in building that case.

From the post:

...bids actually cannot be reasonably be calculated for keywords that are not haven’t already been properly tuned in terms of organization and negatives and match types and ad copy. The data surrounding these keywords is garbage data – putting that into even a very clever bid algorithm or calculation results in a garbage bid suggestion.

Paid search management is a process, not an end goal.  At no point will a PPC program be perfect, nor can it be given the changing landscape outside of our control.  Just because a keyword hasn't been given an ideal (or nearly so) set of "tunings" (negatives/ad copy/etc.) doesn't mean we can't calculate it's current value to the program and save some money for our clients in the process.  As we discover means to improve the performance of that keyword over time we will be able to adjust its bid accordingly.

Craig does have a good point though in that if a keyword is effectively turned off permanently by an overzealous bidding system without ever getting a chance to succeed following optimizations, you are likely limiting the future potential of your program.

For new clients at RKG we will generally advise them that we need to collect data on the order of weeks before our bid management platform can begin fully optimizing bids.  During that period we are also working hard to improve the quality of the program outside of bidding.

For more established programs, testing becomes increasingly critical and it is important to ensure that keywords with changes applied to them get exposure under the new settings.  A keyword may not perform well enough on broad match to guarantee first page placement, but switching it to phrase or exact may just provide enough of a performance boost to justify that minimum bid.  However, a flexible bidding system should allow you to give that keyword a higher bid with minimal effort or even recognize the higher conversion potential automatically, so it is not a prerequisite to determine a keyword's optimal matchtype before applying proper bids for current performance.

From the post:

The idea that paid search success is driven by keywords and bids hasn’t been true for many years...

For most paid search advertisers, bidding isn’t that important.

I think Craig is just being intentionally provocative here as his follow up comments on his post suggest.  Still, I feel compelled to point out that even if you have a nearly perfectly tuned program (outside of bidding) , if you are only running 1/10 or 1/2 of the keywords you should be, you are probably going to take in fewer dollars at the same efficiency than even an averagely tuned program with 2 to 10X your keyword list.  Sure, high keyword counts don't mean much if it is all bloat, but you can't generate a sale from a keyword you don't have.

Ultimately, bid management is critical to the success and scale of a paid search program (it's one of just three components highlighted in the initial graphic on ClickEquations' home page), but it's not the only thing and Craig is right to point that out.  Bids can account for, but cannot overcome bad or broken landing pages, poor query matching by the engines, unappealing and vague copy and a whole host of other issues that may get short shrift if bidding is difficult or given undue attention.

I'm looking forward to Craig's follow up post on this topic and I hope he doesn't mind this friendly response to the first!


7 Responses to "Reports of Bid Management's Death Are an Exaggeration"
Terry Whalen says:
Hi Mark, I run a small PPC agency and we typically see keyword bloat more often than too few keywords. You mention that you can't generate a sale from a keyword that you don't have - but that's not wholly true, since we do have 4 match types, and 3 of those (broad, enhanced broad, and phrase) do allow us to have our ads shown on keywords (search queries) we don't have. And that's why one good process to follow is to keep mining the search query data for negatives and as a source to add keywords, to include visibility and control. I agree that Craig must be writing the post with an eye toward provocation (provacativeness? is that a word?). --Terry
Mark Ballard says:
Good catch Terry, I should have worded that a bit better. To your point, use of non-exact matchtypes is no guarantee you will show as often or with the level of control you would like for certain queries, which is why I agree that keyword research from query logs or other means is invaluable to a PPC program.
Bradd Libby says:
Danuloff coins a great expression: "premature bidding" "Attempting to bid prematurely", he says, "results in inappropriate bids which will *cause* poor performance". As if there's a way to run an ad without first specifying a bid. Here's an idea - Whenever you launch a new keyword, try bidding $0.00 and see how far you get. Obviously, there's no such thing as "premature bidding". I guess he's using the phrase to make people (men, in particular) feel inadequate. You must specify a bid in order for an ad to be run. And then, as performance data arrives, you (or your algorithm) adjust the bid accordingly. So, like everything else in PPC, bidding is an iterative process. That really shouldn't be a surprise to Craig Danuloff or anyone else, for that matter.
Kudos to Craig for stirring the fire, but I agree with Mark that ultimately he's mistaken here. Let's 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. For 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. Should 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, it's also unwise in that you'd end up paying more than the traffic is currently worth with no guarantees that the problems with the ad will be addressed soon. No question, smart bidding isn't the only thing, but it is the surest control we have for making paid search cost-effective.


Check out what others are saying...
[...] This post was mentioned on Twitter by ClickEquations, Chad Summerhill and Metatags, SEM Software. SEM Software said: Reports of Bid Management’s Death Are an Exaggeration: Craig Danuloff just put out a provocative post on the Cli... http://bit.ly/eCCxyd [...]
[...] 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 vitriolic as the dueling headlines might lead you to [...]
[...] take it until some other requirements are complete.What The Pro’s KnowAs several of the smart folks from RimmKaufman pointed out, paid search management is an iterative activity, and search managers can walk and chew gum at the [...]

Leave A Comment