Diorex

So early this week, we noticed on a few of our most profitable terms that our impression volume went through the roof. Like 10x the average daily volume and 8x the highest daily volume in the last 6 months.

On these keywords our clicks went up by approximately 50% of previous levels and our sales volume did not move one iota. Meaning that on 1/3 of our clicks we were getting zero conversion!

So we call up our friendly Yahoo account rep. Who is in an ‘important meeting’ but breaks away to say “fluctuations can and will occur” then says she will look into it. 6 hours later (after we have essentially turned off our account because cost per sale is higher than our total revenue) she dashes off a quick one line email to the tune of “I am having our click quality team (just try and get any of your account reps to say the word fraud, honestly try) look into the matter and they will have an answer within 5-10 days, but our click detection algorithms are the best in the business (another good one, that is like being the best looking girl at the over 300 pound Weight Watcher session).”

Whoa…I was not even terribly concerned with the fraud, I will take the refund, but I just wanted to alert them to the problem of someone figuring out a way of their best in class system.

The next morning we call Yahoo again and get sent directly to voice mail, we then go to her manager who will “call you back as soon as possible”. Finally come late afternoon we call the standard 888-Yahoo-SM phone number and finally speak to someone on our 3rd call who apologizes profusely and goes and gets the manager to call back. The 2nd person we spoke with actually said “that is not my problem” the 1st person said she would get it fixed in 15 minutes and call us back, an hour later the extension she gave us is a dialtone and noone has even heard of her. (Seems like the penalty for trying to help someone out at Yahoo is pretty severe).

So I finally get the manager on the phone. He of course apologizes and says this is not how customer service is supposed to work and that this will be used as a training example, blah blah. Yet he actually says that 5-10 days our rep quoted us is probably the low end of the scale and that with the holiday coming up next week that we probably will not hear anything until afterwards.

He then goes so far as to say, it is funny none of the other ”major” accounts in your vertical have noticed anything. I am tempted to explain why, but decide it does not matter and is not worth the efffort.

The bottom line is that he is totally unable to help us. He is apologetic, but really does not seem to care.

So we are left to our own devices. We ended up putting their Yahoo analytics on a single high volume keyword to track referrers. We accidentally forgot to properly install the conversion pixel, because that is none of their business. Less than an hour later, we had a single referrer that had more than double the volume from Yahoo itself, without so much as a click through on that page.

That does not seem like something that should have actually been rocket science for Yahoo to figure out. 10x impressions was the first start, how hard is it to isolate impression volume by IP and flag suspicious volume, especially when an advertiser brings your attention to it.

So based upon our experience with Yahoo who did not think there was a problem, we decided not to alert them to the non-problem url, we figure in a few days the ‘major players’ who don’t monitor their account nearly as closely as we do will see the decrease in conversion and raise bloody hell. Our result was almost immediate, conversion rate recovered and impressions fell back to normal levels. Problem solved, except for the problem.

Paul at Uber Affiliate shared a tip with me about GeoTargeting at Ad:tech and then blogged about it today.

We had tested this in the past and our results were that our CTR went down significantly and our CR also declined slightly. One thought for the change, is that we are selling national products that people probably do not think of as state specific. I was more profitable even with the non-targeted out of zone clicks (which was less than 5% of total click volume) so since my primary goal is total profit, I preferred to do without the Geo-Targeting.

A second thing is that Google’s geo-targeting is screwy. For a long time our search results at the office thought we were in a city and state more than 1000 miles from our actual location. I have spoken with other industry Pro’s who also have similar problems. Also dont forget about the school district that sued Google because it showed ads for another country. I don’t exactly get a warm fuzzy when an advertisers tells me I can get a great deal 4 states away - their credibility (courtesy of Google) is shot from the start.

I think Paul’s idea of testing this is very valid and will probably work for some verticals. The problem is that lots of people will just read and implement without testing. Anything you read on my blog or any other should be taken with a grain of salt and only implemented if it makes sense for your business, as well as actually performs better than what you have now. The only way you will figure it out is by testing.

Just read about how Google is allowing advertisers to exclude different types of traffic such as Error Page or Domain Ads, as well as social networking, tragedy, gross-out and other pages.

I applaud that they are finally allowing advertisers more control over where there ads might show. I am frustrated that the additions are all or nothing.

If Google really wanted to let me have control they would let me run all of this stuff independently with bids and tracking for each. Say I just wanted to bid on error page ads, why not let me. As it is set-up now, I cannot put different bids on error page ads than I have on search ads.

When they let me do this in Content, my spending for that segment went from $0 to several million a year.

I suspect that domain ads convert really well (ok I know they do, but not all of them - negative site match domainsponsor.com and enjoy the increase in conversion) and might even be willing to pay more for them than I do for search…

What about the search network? I want to buy AOL traffic - I am lukewarm about Ask - I want to avoid Iwon and Lycos like the Plague. My search network bids end up being an amalgam of conversion. Say 80% of the traffic and 90% of the conversion comes from Google. I end up bidding less in Google in order to offset the lower conversion elsewhere. So I end up paying Google less for the clicks where they get 100% of the revenue.

Bottom line - the more segmentation and options you give sophisticated buyers, the more they will spend with you. I will spend it in different ways and in different places within the network based upon my conversions, but in the long run it will help Google earn more from me. (A scary thought…)

While I am just starting to get into the domain game with some personal investments, I am a sizable buyer of this traffic through various search engines and other products. Earlier this week Frank Schilling wrote about traffic quality. For those of you who do not know Frank, he is very insightful and his blog is well worth your time and effort.

In this post he talks about how the potential for good traffic sources to have their CPCs adjusted lower because they are lumped in with lots of poor quality traffic.

I am in the process of buying lots of names and parking them while I wait to either develop or sell them. I do not have much traffic, but most of the niches are of high enough quality as to be able to pay the annual domain bill with just one or two clicks.

I have now tried 3 different services and am finding the value of similar clicks varies wildly. I am not a large enough traffic volume to get in to any of the “elite” programs so I am just left with what will take me.

As a sizable buyer of this traffic, I noticed one of the better known and I assume well respected players in parking start to show up in my Google content reports no longer as Parked Pages, but actually start to show up as their own name. Not sure why this happened, but it did happen about 3 weeks back.

Assuming that most parked traffic is valuable, I asked one of my analysts to take a close look at conversion rate with my eye on trying to site target or otherwise increase my volume from this player. What we learned shocked me. The CR from this one traffic partner which was 5% of that accounts traffic, was 0.6% at a CPC almost 30% higher than my average.

The traffic quality was horrible and it was expensive as well. I would have found this in our end of month report, but I was so excited at finding the potential ability to exploit an opportunity that I expedited the process. Turns out this one partner was creating a drag of about $1200 a day in profit from this poor traffic quality. I quickly checked out the data on my other content accounts and learned that this was not an isolated event. I was losing money on this parking company on every single account I have, even some of my personal stuff.

Now finding poor converting partners and turning them off is not that unusual. What is unusual is that this particular parking company not only took the initiative to make sure that Google reports broke out their traffic, but they did not seem to care that they were mostly representing junk traffic. My conversion rate was so radically different (and the traffic volume pretty high) that this was very likely click fraud on some level.

If you are a legitimate domain owner and just happened to have one of your domains sending me traffic, you have just lost out on about 3 of the top 7 names in that industry, to be replaced with much lower paying arbitrage type traffic. Their was obviously some legitimate traffic that if isolated would probably be profitable to us, but when surrounded by all of the bad company we cannot afford to buy your traffic any longer.

I am sure some domain parking companies are very vigilant about traffic quality and very protective of their reputations, those should allow Google to isolate them so sophisticated traffic buyers can find and bid directly at your high quality traffic. While others seem to sell my Google ad results to any 2 bit punk with a dial-up connection that thinks they can have friends click on them and make some money.

This is a 2 way street and I now find myself to be standing in the middle of it. I am willing to pay a very high and fair wage to traffic sellers of good quality, while I am unwilling to even think about bidding on poor quality traffic. At the same time, I have found a few domains that get enough traffic to be worth parking them, the traffic is typed in generics so likely decent quality since that person has just stated that they are looking for that service by typing it in. Yet I wonder if I am hanging out in a bad neighborhood without even knowing it.

In the future I expect to see far more visibility from all major sellers of traffic which will allow buyers to pick and choose based upon quality.

As an advertiser, are you buying quality traffic and not buying bad traffic? Do you take the time to find out?

As a domainer, are you making sure to surround yourself with only other quality sites? The kicker here is how can you even tell?

Hey guys - a little help please.

Yahoo is rolling out the ability to Block Domains in the near future. The only problem is that we are having a hard time getting visibility into the domains we might want to block.

Google has long had the &site={placement} tag that will pass the referring domain so we could track conversions.

Something Yahoo does in their redirects keeps us from knowing where a huge percentage of the Yahoo traffic originates from. I’m not the servers guy and I did not understand the Swahecian Tech dialect he spoke trying to explain the issue.

We have of course asked our Yahoo reps who are normally pretty helpful, but they seem to be clueless here.

Does anyone know of a tag we can put into the landing page URL that will allow us to collect the referring page in Yahoo?

Ok here is something else I have been playing with that is having mixed results.

I find a page for which I want my ads to appear, but they are not currently. I want to be CPC, but I dont want the ugly site targeting ads that cost a fortune, I just want to show up in the normal adsense ads on that page.

This is probably stupid simple but here is what I do…

Go to the Google Tools page
Click on Keyword Tool
Then click on left tab “Site Related Keywords”
Then I copy and paste the URL where I want to appear
Make sure the checkbox is unchecked
Copy and paste all of the keywords Google suggests into a brand new adgroup

I usually not only show on that site, but on others I had not been aware of previously.

Not full-proof and not always cheap, especially if you are bidding on keywords not related to your site.

PS. Want to know what keywords Google thinks your landing page is about, do the exact same thing with your URL, use that list to massage your quality score for search.

I want to thank Sup3rnova for the mentions in his Uber Affiliate Marketing Guide. Best of all it has a low price of just $97, but if you click now there is an instant rebate for my readers of $97. Thats right he has compiled everything you would ever need to write an Ebook on making money in internet marketing totally free!

My traffic has skyrocketed as a result of his post, which I appreciate. But, I appreciate that he took the time to find all of these gems. I actually ended up re-reading some of my old stuff and all of the other links in there. Some absolutely amazing blogs I have never even heard of, as well as tons of great content, tips, secrets, and commonsense.

It is free, but it would be worth every bit of the $97 he should have charged…

Assuming that you have a somewhat mature PPC lead based (not products) campaign within a vertical that has all appropriate keywords with properly set bids etc. I can think of three simple ways to earn more profit from that campaign.

1. Pay less for each visitor - most people attempt this by lowering bids dramatically - this seems to be the most common and least profitable, since you typically also sacrifice volume. You can also improve your quality score (think SEO on landing pages) hoping to get a reduction, but this is sometimes problematic and certainly not a slam dunk process. Another idea is to find areas where your max bid is significantly higher than your actual CPC. Say you are bidding $3.00 and the CPC is $1.75 - try lowering the bid to $2.50. 9 times out of 10 your CPC will go down without a change in position.

2. Convert more of the existing visitors you have - Optimize your landing page to convert at a higher pace. Change things subtly (button text to “order now”, or “Continue” or “next”) or dramatically (New hero shots, totally different layouts etc) and take the same pool of visitors you have now and convert more of them into sales. If you control the form, be sure to optimize that as well. Since you have already paid for the visitors each additional sale is 100% pure profit. Small tweaks can result in huge gains.

3. Get paid more for each sale - Go directly to the vendor and ask for higher payouts. Cut out affiliate network middlemen. Do not allow others to profit off of your work. If they will pay $25 a lead, they will probably pay $30, especially if you turn the faucet off for a day or two.

Simple Numbers to illustrate this point (totally made up…)
Assume 1000 visitors a day at a $1.00 CPC with $2.00 max bids.
5% of those visitors convert = 50 sales.
Payout per sale/lead = $35

So you are making $750 per day from this campaign (($35*50)-($1*1000))

Now assume you:
+decrease CPC by $.10 a click without losing position by reducing bids to $1.50 for a CPC of $.90
+Increase Conversion rate to 6% = 60 sales
+Increase Payout to $40

You are now earning $1500 a day (($40*60)-($.9*1000))

3 relatively small improvements doubled your daily profit.

Once you get something that is modestly successful, really small tweaks have dramatic consequences.

****There is a 4th lever which many fail to consider which is that you can increase your bids, move up in position, get more clicks and earn less per click, but probably improve conversion rate as well, for more total profit. This is risky and not for the feint of heart.****

First in a series responding to readers requests (if you have any requests/suggestions for discussion topics, please say so), lets talk about keywords.

First off, my philosophy on keywords is that anything can be valid as long as it converts at an appropriate rate to be profitable. So test even things that are just tangentially related to your field, who knows there might be a goldmine you have not yet discovered. If you are not tracking to the keyword level, then you are not paying very close attention. Tracking should be a major focus.

Second - match types. I start just about everything on broad match and then move the stuff that works to [Exact] match. I dont use “phrase” matching at all. I used to, but found it typically did not add incremental revenues. I set-up a report in Google to be sent at the first day of every month to show me all the match types from the previous month. I typically take the top 20-50 (depending on click volume) and make those exact match keywords.

Third - Number of keywords - I dont really care about many more than my top 30-50 keywords. Over time that number has grown steadily but it still relatively small. With spreadsheets and Adwords Editor there is almost no incremental cost to have 10,000 keywords vs just a few dozen. Keep the ones that work and kill the ones that do not work.

Fourth - negative keywords - use the monthly report from above to filter out things you dont want or that do not convert. I cannot sell my product anywhere other than the US. So searches for UK or Canadian or Austrailian “keyword” are wasted and I dont want to pay for them. Same thing with credit cards - you have to have one for me to get paid, so I negative match keywords like “no credit card” etc. Lastly there are some search terms that just will not ever convert regularly like “customer service” and “scam”. Get rid of what does not work for your over time. Save just a few dollars a day and that will add up to money in your pocket.

Five - iterations - dont forget plurals, misspellings, competitors brand names, URL’s and other things that will have lots of different ways to be searched for. (think www keyword com vs keyword.com vs keyword com and about 20 other URL variations some are obvious some will be seen through the monthly report)

Six - new keywords - I no longer pay for any keyword tools, they all pretty much were some variant of the overture tool or otherwise flawed in data provided. The overture tool is no longer anywhere as near as useful as it used to be. lately, the way I get keywords is to type in the premium keywords in a vertical and see the top 5-10 URL’s. I then go into the Google keyword suggestion tool and click on site-related keywords. Type in the URL’s of the big guys and be sure to check the box Include other pages linked from this site. Then just scroll to the bottom select all, rinse and repeat and save into a CSV. then wash against the keywords I already have in Adwords Editor and then hand check to make sure to eliminate the obviously wrong keywords. I can usually add 500-1000 keywords every time I do this. As Google’s tool learns it produces better results, so be sure to come back once a month. One account earns $500 a day from my Google Suggestions AdGroup and that excludes the keywords I have graduated out of there into other adgroups. Ignore Google’s volume, CPC and position guesses they are not even close both high and low.

Seven - Organization - I have Campaigns for broad match, exact match, misspelling, and Misc (google suggestions etc). In this way I can easily locate what I am looking for. Content always goes into its own Campaign. I dont mind putting related words in the same ad group, but anything getting any traction in the way of clicks gets put into a 1 keyword adgroup.

Eight - Match ad copy and landing pages to key ad groups - you can make subtle changes to landing pages that will barely touch conversion but that might significantly improve quality score and thus CPC. Try putting the keyword in the landing page title (not dynamically) or in the URL (either subdomains or interior URLs), or work it into the ad copy, even paying attention to plurals can make huge differences. Anymore, the little details are becoming more and more important.

That is everything I can think of in relation to keywords - a lot of rehashing and my personal opinion - your results may vary.

Recently launched a new campaign in Google for search only. A brand new URL that was recently purchased and being used in Google only with Content network turned off.

I did a search for the URL string in the Google results to see if the page had been indexed yet.

Strangely, there were 2 results. One for my domain and another for the domain ChamberofCommerce.com (slimeball is good anchor text methinks) which has nothing to do with the US Chamber of Commerce or any other local COC for that matter.

So I went to the page (I am not showing up on the page I linked to…) and found a long list of sponsored results for which my ad was showing up amongst them. No Ads by Google link or anything. Just some scraped content (everything I copied and pasted from the site I found on other scraper type sites…) and then a huge list of sponsored results…

I had never seen a Google Ad group with 10 ads so this immediately got my attention, since not only are my ads only running on the Google network, but they are not running on content..So I investigated further…

I right clicked on the ad and clicked properties:

hotkeys.com/leadtracking?id=…ask.com (appended so it fits on the scree) which was then followed by a huge string of numbers….

Ask.com…I don’t advertise on Ask and never plan too after several experiences where they were distributing my ads outside of the search traffic, through Clicksor ads. I am a little pissed at Google that they will syndicate my ads and then allow that entity to re-syndicate them.

I immediately contacted my Google rep and explained the situation. They investigated and responded that this ad was not being served by Google, rather that it was being served by Akamai, which was not at all affiliated with Google.

Google was adamant that this was not a Google ad despite the fact that it landed on the proper URL, even going so far as saying it was possible someone was running this and not charging me….

What I had not told Google was that since I had rotating ad copy, that my ad was showing up differently each time I hit the refresh button. I asked the rep to hit refresh and the response was “we are not serving this ad, but we will see if there is some way that someone could know to rotate your ads.”

Even with pretty convincing proof of Ask.com serving my Search Only ad in a content setting, Google was unable to accept it. Now Ask is probably more important to their business than I am (and who am I kidding, I am not gonna quit Google…) but even when shown a blatant example of them intentionally re-syndicating ads in an improper manner they just do not seem to care.

Overture feeds are the single biggest problem at Yahoo. Google’s largest issue is the parked pages and that they seem to think the quality of traffic from excite, ask, & myway is the same quality as you see from AOL or even the Google network itself. Yes, I know you can go Google only, but I have tested it and even with all the fraud and junk clicks I am still better off than otherwise.

Google - you recently gave us insight into Content, but I think you need to follow that up and give us insight into the search network. Let me know where my ads show, give me a way to track it without having to use Google Analytics, and allow me to opt out of anything in your network, or set proper bids for individual traffic sources and I will pay you far more than I do now, because I will be able to weed out what does not work and concentrate on getting as much of what does work as possible.

Unfortunately, search engines seem to be in a race to the bottom, and quality suffers. They are falling all over themselves to give syndication deals to crappy engines with huge fraud potential. Google states they want to be about quality user experience, but seems to fail to recognize that the most important users are those footing the bill, what about the advertisers experience?