Forum Moderators: buckworks & skibum

Message Too Old, No Replies

Update - PPC Profit Maximizing (not ROI targeting) bid management tool

Conclusions from older thread by enterthevortex

         

enterthevortex

12:06 am on Oct 1, 2008 (gmt 0)

10+ Year Member



This is an update for this thread [webmasterworld.com...]
The reply function was disabled, so I posted it as a new thread.

Hi all, for a while this post was ripped down so I didn’t know it was still going. Here’s an update of my experience.

Searching for a third party solution for my needs has been a long, fruitless, aggravating experience. I know you were all hoping I’d be the guy who would find the solution, but I’m sorry to report that I did not find anything worthwhile.

First, I gave up on finding anything more sophisticated than target ROI bidding. If you go back to the initial post, you’ll read that I was looking for a bidding system that in some way adjusted for profit maximization and that ROI is only nominally relevant to profit maximization. The thing is, bidding to ROI is not very complicated unless you are managing 100K+ keywords, so I don’t know why people with deep pockets and high ad spend would pay an agency to do this for them. The answer is that they probably don’t know what they are doing or don’t want to take development time, so they just hire an agency (that’s fine, these agencies fulfill a need in that regard).

Second, I didn’t find something that managed ROI bidding at an acceptable level for my specific purpose. (I know what I said in the previous paragraph, but I figured that a third party must have come up with something that could outperform what we would create and justify fees through performance & service.) Realize that bidding solutions must be tailored on a case-by-case basis to fit your needs. Everything I saw from these agencies was like jamming a square peg into a round hole. We kept asking ourselves “how can we make this work?” We should have been asking “why should we pay x% of our ad spend when they don’t provide a solution, just a different form of work?”

In my last post I stated that I was going to try Marin and I was hopeful from the phone sales calls that they’d have a solution. That experience was painful. We were clear up front about our need for making decisions on small sample sizes of data. They explained their solution in statistics language that made perfect sense to us. Problem was, they didn’t really have that solution. We lost months of time waiting around while they stalled. The customer service was great in the first couple weeks, but it quickly deteriorated and communication became frustrating. Here’s a drinking game if you ever deal with an agency… Every time you hear “roadmap” take a shot – caution, seek immediate medical help after phone call.

So what did I do next? We set out to build our own customized solution. We decided to settle on ROI targeting for simplicity. I won’t say it’s been going great at the keyword level, but much better than our experience with third parties. Our biggest challenge is dealing with the massive number of keywords we bid on. If we dealt with only 50K keywords or so, we’d have a keyword level bidding system with ROI targeting that fit our needs. But pulling and crunching the data becomes much more complicated with large data sets. We have adgroup level ROI targeting solved (for many thousand adgroups), but not keyword level yet.

In conclusion, we didn’t try every agency out there – in fact we only tried two. However, we took sales calls with all the major agencies that looked promising. We were able to rule most of them out based on an inability to fulfill our specific needs. I’ve read everything since my last post. To be honest, I don’t have the energy to give it another round on sales calls. I read that shorebreak is working at Omniture now. I remember Omniture wanting me to sign an NDA before discussing some details. That’s some bureaucracy that I don’t want to jump through on a sales call.

Sorry guys, I have no answers for you.

jskrewson

3:16 pm on Oct 1, 2008 (gmt 0)

10+ Year Member



I've been working on a home-grown, keyword level ROI maximizing software program for years. This program is intended to monitor and adjust 1 million plus keyword campaigns. I've come to the conclusion that its best to maximize ROI at a higher level, e.g. adgroup or campaign and not worry about the profitability of individual keywords unless they are the outliers. Here is my reasoning...

1) Longtail keywords, while profitable, don't have enough statistical sampling to make an informed decision.

2) Outside factors too frequently affect ROI for a particular keyword. It's like the warning that comes with mutual funds. Past performance doesn't give an indication of future returns.

To make things even more complicated, I'm running ad campaigns for consumer products and their lifespan is sometimes only six months. So as soon as I start to get great statistical data on the adgroup or campaign level, boom, the product is obsolete.

Honestly, now a days, I use a computer program in conjunction with human intervention. The software identifies possible problem areas then a human goes in, does the research and figures out what should be done. Its a huge pain in the butt and I wish I knew how to outsource the entire process.

jskrewson

3:25 pm on Oct 1, 2008 (gmt 0)

10+ Year Member



I've also found that truly maximizing ROI reduces the amount of traffic my site receives to an unacceptable level. I actually aim for a balance, e.g. fairly heavy traffic and decent ROI.