Jul 92009

PPC Bid Management Misunderstanding #1: Manual Bidding vs Automation

There's an awful lot of hooey out there regarding bid management systems. Some of it is a product of limited experience ("automated bidding wastes money" claimed one recent and thoroughly misguided article). Some of it comes from irresponsible marketing blather promising to make wine from water.

However, sometimes the arguments stem from genuine confusion over terminology and a failure to recognize that different strategies may make sense for different types of advertisers.

This is the first of a multi-part series attempting to delineate these issues.

Manual Can Be the Right Approach

Setting bids by hand or relying on the engines' own bid management tools can make sense for many advertisers. Indeed, it probably makes sense for most advertisers by head count. There are two very good reasons to eschew automated bidding:

  1. There isn't enough data to make automation worth the cost. 2/3 of Google's advertisers spend less than $5K per month.
    • When the data is thin, algorithms struggle to make accurate predictions. The signal to noise ratio is too low.

      In fact, using Google's own bidding system can help alleviate this problem. Google's system uses conversion data aggregated across every company using Google Analytics conversion tracking. If your data is too thin they can guess at conversion rates based on others advertising on the same keywords.

      Using average conversion rates across multiple advertisers is an awful idea in general. Conversion rates and AOVs vary based on price, selection, brand familiarity, and site usability. However, if you have nothing better to go on, it can be better than nothing.

    • Also, if the volume of data coming in is thin, the need for and benefits of rapid responses to fluctuations in value vanishes.
  2. There isn't enough complexity in the program to justify the cost of an automated system. Simplicity can come from a variety of different sources:
    • The number of keywords needed to thoroughly cover the offerings is small. If there's no tail or midsection to speak of and only 100 keywords that matter, managing bids by hand is fairly quick and simple.
    • You're a manufacturer and the only keywords you compete on are your brand names.* If you're really only competing against your distributors you should be able to outbid them on your trademarks as you have more margin. It doesn't take a fancy algorithm to bid to the top of the page.* Manufacturers often have trouble competing with distributors on non-brand competitive search because their distributors carry their products in addition to similar products from competing manufacturers. The extensive selection gives the distributor a huge conversion rate advantage on general, product category/ sub-category related terms and can leave the manufacturer with nothing but their trademarked keywords driving traffic.
    • There's no competition. If you are all by yourself on the page, bidding is pretty simple.

Automated Bidding is Essential for Others

Calculating and setting bids by hand leaves tremendous opportunity on the table for large, complex programs. The cost of automation vanishes when the programs get to be large. Companies spending more than $15K a month and managing thousands of keywords by hand are missing opportunities. There is no way to keep up with changes in performance owing to seasonality, the economy, the time of day, day of week, offers etc. across a wide array of categories.

For the first year we were in business (October 2003 through 2004), we had a smart algorithm for doing bid management but no API interface through which to automate the updates. We updated bids smartly, but roughly once a week. Once the API interface was in place and we could update bids as often as we liked the performance of our client's programs improved markedly and immediately. We're not guessing that it matters, we know it matters.

However, all automated bidding systems are not created equally. Setting bids smartly but infrequently by hand is certainly preferable to setting bids stupidly but frequently by machine.

In the next post I'll talk about different approaches companies take algorithmically, and the confusing use of language that can make it difficult to tell who's doing what.

Bid Management Misconception #2: "Bidding Rules"


13 Responses to "PPC Bid Management Misunderstanding #1: Manual Bidding vs Automation"
Mel66 says:
Spot on, George. I love all your articles but this one is gold. "Setting bids smartly but infrequently by hand is certainly preferable to setting bids stupidly but frequently by machine." Awesome. Great stuff.
Thanks Melissa! We've had the opportunity to take over programs from big agencies, small agencies, affiliate networks, in-house managers, you name it. The programs we've taken over have ranged in quality from absolutely horrendous/ "someone should face charges over this," to pretty solid. The differentiators between horrid and solid is not automation, but algorithm and whether they're thinking about search the right way, paying attention to the numbers, and reacting correctly to them. Some of the absolute worst programs have come from big agencies/networks that most certainly have automated tools.
George brings up all valid points. However, keep in mind that "automated" bidding still requires an analyst with a brain to closely monitor a complex paid search program. The more data the better to build out models and prediction to help achieve the client's acquisition targets. Automation isn't a set it and forget it solution.
Kenny says:
George, I read that article the other day too and gasped. Terribly misleading - venting because you haven't found the right model that works for your business yet doesn't mean that "automated bidding wastes money". I'm glad, though, that you provide balance in this post too, because I have seen just as many if not more articles claiming that bid automation IS ppc management. If you don't understand the more "artistic" aspects like account structure, keyword choices, match types, and Quality Score factors, a scientific approach to bidding may help but you undoubtedly leave massive amounts of money on the table.
Excellent point, Jeff, thanks for bringing it up! In fact you've anticipated my next post in the series!
Thanks Kenny, You're right, there are many many pieces to the puzzle that need to be handled correctly. I'd rank the bidding algorithm near the top of the list in importance, probably just behind Keyword - landing page combinations, but the mechanism for implementing the bid changes recommended, and the frequency are less important than getting the numbers right. One big advantage of automation in general is that it frees up time for analysts to spend analyzing data, building keywords, testing copy, etc. The less time people spend doing robotic tasks the more time they have for other high priority stuff.
Kenny says:
Absolutely agree, George. Once accounts get to a certain size in particular, automation is desperately needed so that account managers/analysts can spend their time mining the data looking for nuggets of further optimization gold. My personal experience has been that quite often bidding changes in terms of a better algorithm or increased frequency of bid changes undoubtedly improve ROI, but sometimes in minor steps up, whereas optimizing other core campaign related items mined from data regarding keywords/landing pages/etc etc can step ROI up in large chunks.
Really interesting, Kenny. Certainly we find programs that are broken in a variety of different ways, and we can usually anticipate the problems based on who managed the program before us. Some firms under-appreciate the importance of KW and Landing Pages, and we know what we're going to find when we look under the hood. Others put reasonable thought into the foundations but have no concept of analytics and bid management. It seems to me that we've seen a shift over the last couple of years from more programs getting the most benefits from KW buildout and landing page connections to more programs missing the boat primarily on analysis/bidding. Possible that there is a selection bias at play. It could be that more firms are coming to us because of our bid management approach so the balance seems skewed that way.
Andrew says:
This is a great post. I was intrigued when he mentioned how the programs suffer with big agency work. It seems as though with some large agencies the work is "factory formed" and not much tailoring and attention to detail is done. Anyway, If you would like to give some feedback on how to improve programs of this nature click here. http://www.loyaltyleaders.org/survey.php
Hi George, One of my main concerns with rules-based bidding though is that it can create a false sense of security for the search marketer. The "keywords are now being taken care of, so I can now forget about them and do something else" mentality can be hard to avoid when there's a robot doing all the work for you. In my experience, the whole process of making manual bid changes helps me relate to the account and better understand each keyword's relative importance. It's often when I'm actually making bid changes that I think of new ideas for keywords and ads.


Check out what others are saying...
[...] we can always rely on RKG’s George Michie to provide a balanced and sensible discussion. George is one of the best writers when it comes to PPC and I read his blog avidly. The most recent [...]
[...] with the question of Automated Bidding, it turns out both of these folks might be right depending on what they mean by “bid [...]
[...] PPC Bid Management Misunderstanding #1: Manual Bidding vs Automation www.rimmkaufman.com/rkgblog/2009/ – view page – cached PPC advertising is not a one size fits all game. There are cases in which manual bidding makes sense. There are cases in which automated bidding is essential. Here's our take on which is which. — From the page [...]

Leave A Comment