Archive for the ‘CPVLab’ Category

About 10 days or so ago, I wrote this CPVLab writeup. For full disclosure, I mentioned that the review was based solely on sales video, discussion with the creator, and testimonials from other users.

At that time, I wasn’t using it. Now, it’s about a week and a half later, so what’s the PPC.bz verdict?

CPVLab – Yay! or Nay?

CPVLab Gets the PPC.bz Seal of Approval

That’s gonna be a ‘hells yes!’ The software is awesome, plain and simple.

It’s not some miracle cure for making monies. You still have to know what you’re doing – how to find targets, how to write copy, how to make delicious landers, and how to convert. I just want to cut through any misconceived notion that CPVLab is some “push a button and make monies” guru product. It’s not.

CPVLab is a campaign management and tracking system. If you don’t have campaigns to track and optimize, then it’s of no use to you. If you do have campaigns to track and optimize, or know where to start, then it’s a great tool to have in your toolbox. It will save you time and money, no doubt.

Even after 10 days of using CPVLab, I still haven’t touched every feature. There are 6 campaign types you can set-up, and I’ve just been sticking with 1 or 2- Regular Direct Linking / Landing Page campaigns. You can get into stuff like Lead Capture and E-mail Follow-up, but since I haven’t used those yet so I won’t get into them for this post.

Now onto the stuff that has impressed me the most.

On-the-Fly Campaign Management

This is probably my favorite feature so far. Campaign set-up is a breeze. Just throw in your URLs to your pages, put the LP code on them, then go.

But once a campaign is up and running, the ability to modify stuff quickly is fucking clutch.

Adding New Landing Pages – If you’re a 202 user, then you’re probably using the LP rotator script. The steps involved when adding a new landing page would be – 1) Open your php file, once you navigate to it in File Explorer, Dreamweaver, or what have you 2) Modify your landing page rotation array, where you have to do some match to make sure your % delivery is correct 3) Save 4) Open FTP program 5) Upload (or SVN it).

If you had 3 campaigns going and wanted to add your new landing page to all 3, then it starts to get a little tedious.

The steps to do this in CPVLab are 1) Paste in URL to your campaign 2) Hit Save. Donezo.

By the way, if you’re looking for a 202 rotator script with cookie functionality, I made one here. I make no guarantees on the functionality of this script. Use and test at your own risk.

Example: I tossed up a new campaign a few days ago with some fresh landing pages. I did a small change to one, and used the % Share Distribution feature in CPVLab to split them 50% / 50%. After an hour or two, the data started to show both of these landing pages were doing pretty shitty. Both had CTRs under 1.5%, which is unacceptable for PPV.

‘Fuck me’ I thought. Not to worry, I have CPVLab!

I took an old landing page I didn’t think would work, and tossed it up anyway for shits and giggles. I set the distribution on the new one to 70% to dedicate most of the traffic there, since the other two were losers. The losers still had 15% share each to give them a chance- just in case it was a fluke.

Also, thanks to the alerts which highlight any target not performing to your specifications, I saw that 2 targets were getting high traffic but giving me awful CTR.

With the quickness, I paused those two targets and had the new Landing Page getting the majority of the traffic. A few hours passed by, and the new landing page hovered around 7 to 9%. The 2 losers improved slightly because I dropped the bad targets, but never broke past 2.5%.

With just a few hours of data, the campaign is profitable. I set the winning landing page to 100% to suck up all the traffic until I revisit the campaign with new landers and changes.

This is the simplest example of getting PPV with landing pages to work. No bullshit really. Throw up a campaign, wait a few, trim the bad targets, and trim the lower-performing landing pages. Repeat process over and over until you’re only seeing marginal gains. It’s just so damn fast doing it with CPVLab.

Pro Tip: Running a Test Properly

Some marketers, specifically newbies, do their split testing incorrectly. I’ve definitely fucked up plenty of tests in my day too. They 1) dedicate too much traffic to new pages, sucking away money from their profitable pages. Or they 2) test too much at once, and make decisions with insufficient data.

1. Proper Testing Distribution

– Let’s say you have a profitable campaign going. You want to test a flashy button on your page. You create the new page, everything equal expect that new button, and distribute traffic 50% / 50%.

While it sounds like a good idea, it’s basically a coin flip. During the test, if the new page does worse, you’re going to lose money. If it does better, then you’ll make some money.

Pre-Test
100%: Landing Page 01 = 50% ROI
During-Test
50%: Landing Page 01 = 50% ROI
50%: Landing Page 02 = 20% ROI
Post-Test
Let’s say you spend $200 a day on this campaign. You run the test for 10 days. Total spend would be $2,000.

Before the test, you’d make $1,000 profit over those 10 days ($2,000 Spend x 50% ROI = $1,000 Profit)

During the test, you’d make only $700 profit over those 10 days ($1,000 Spend to Page 01 x 50% ROI = $500 and $1,000 Spend to Page 02 x 20% = $200. $500 + $200 = $700.)

You just shit the bed on $300. This can be particularly devastating if the new page produces negative ROI. Granted, most people wouldn’t let a losing page run this long, but let’s say they do in this example.

Even Worse….
What if you want to test two variations of this flashy button? You split 3 pages, your control and your 2 experiments.

33%: Landing Page 01 = 50% ROI
33%: Landing Page 02Green = 20% ROI
33%: Landing Page 02Red = 25% ROI

For the sake of me not having to bust out a calculator, let’s say you spend $300 a day on this campaign over 10 days. $3,000 ad spend.

Before the test, you’d make $1,500 profit over those 10 days ($3,000 Spend x 50% ROI = $1,500 Profit)

During the test, you’d make only $950 profit over those 10 days ($1,000 Spend to Page 01 x 50% ROI = $500 and $1,000 Spend to Page 02 Green x 20% = $200. $1,000 Spend to Page 03 Red x 25% ROI = $250. $500 + $200 + $250 = $950.)

The bottom line in this example is: Your flashy button does not perform better than your control, so don’t be a fuck-up and dedicate 66% of your traffic to the flashy button! Basically you lost $550 because of your shitty testing skills. If your 2 new pages did really, really bad, then you’d be losing money hand over fist until you realized your mistake.

Conclusion: The idea is to not wrap up too much traffic in untested waters. Whenever you run a new test on a landing page, keep the distribution to a level that won’t ruin your numbers. Basically, MINIMIZE YOUR RISK!

Edit: After writing this post, I recalled that a fellow named CPA spelled out some similar advice about proper testing.

Keep your control as the control! It’s science. It’s why its called a control.

If you have one new page, you’d split it 80%/20% to your Control and Experiment. If you had two new pages (adding two variations of a flashy button), you’d split it 80%/10%/10%. Most goes to your control, and a bit goes to each Experiment. (Btw, 80% is not the magic number or anything. Use whatever numbers you feel comfortable with.)

This is a specific pitfall with 202′s landing page rotator script. The code looks like this for splitting 3 pages:
$landingPage[1]=page1.php;
$landingPage[2]=page2a.php;
$landingPage[3]=page2b.php;

It looks correct and works correctly, but its still a 33% distribution. If you want proper distribution, your code would look like this (assume 80/10/10)

$landingPage[1]=page1.php;
$landingPage[2]=page1.php;
$landingPage[3]=page1.php;
$landingPage[3]=page1.php;
$landingPage[4]=page1.php;
$landingPage[5]=page1.php;
$landingPage[6]=page1.php;
$landingPage[7]=page1.php;
$landingPage[8]=page1.php;
$landingPage[9]=page2a.php;
$landingPage[10]=page2b.php;

I’m not sure how many people fuck this up, but I think its a good amount. Maybe people don’t realize its wrong, or maybe people don’t like doing math. Either way, it can fuck a campaign up if your new experiments are performing worse than your control.

Either way, CPVLab makes this a breeze. Just type in % numbers in your campaign. No need to fuck with your landing page array in a php file. You can edit these %’s on the fly. If something is working better right off the bat, login to your Control Panel and up the % a little bit, or all the way. Make $$$ ho.

Note for CPVLab Users – Just because it’s CPVLab doesn’t mean you’re immune to bad testing either. You can hit the ” = ” button to distribute the % evenly and think its good, but its not. You have to do the math in your head and input the correct numbers. The control would be 80% and the tests would be 10% each, or whatever numbers you choose.

If you start a new campaign with 4 completely different landing pages, then 25% to each is fine.

But if you start a new campaign 2 different landing pages, while rotating 3 different images on one of them, then 25% each is not fine. You want even distribution to different styles of landing pages, not the individual changes on a single page. The distribution on this example would be:
50% to Page01
17% to Page02-w-Baby-Baby
17% to Page02-w-Baby-Girl
16% to Page02-w-Mommy

2. Making Decisions with Insufficient Data

This one is a killer for many people starting out. I’ve definitely fucked this one up and continue to do so sometimes. You here the catch-phrase “Test Everything” thrown around like its going out of style. If you’re a newbie, you might take this to heart and actually test everything.

The problem with testing everything is that each change requires sufficient data to make good decisions. If you test 4 different styles of landing pages, 4 pictures one each, and 4 Call-to-Actions on your buttons, that’s 64 variations you are testing. (This an extreme example because no one is setting up 64 variations with 202)

If you set your decision-making at 1,000 views on a page, that means that this test will cost you $640 (64,000 views @ $10 CPM) to see satisfactory results. I don’t know many people starting out willing to spend that much on their affiliate marketing endeavors, let alone on a single campaign (even though that’s what this game is)

Most newbies will run this kind of test, but bitch out at after spending $100. Let’s say only about 6 pages will work out of the 64. 10% of the pages in your experiment make money, 90% don’t. After spending $100, $90 went to unprofitable pages and $10 went to profitable ones. No matter which way you slice it, this is not enough data to make any kind of decision whatsoever.

Most people just give up on the campaign entirely at this point and call it a bust (No shit when you only sent $10 worth of traffic to pages making money) They then blame shoemoney for selling them the dream when the reality is affiliate marketing is one cold-hearted bitch.

Trying to test many variables at once is solved with multi-variate testing, but its not something I am an expert on, so I won’t ramble on about this shit anymore.

The point is that you only need to test the important stuff first. The headline is the most important part. I wrote about the importance of headlines here.

You don’t have to TEST everything. Just TRACK everything so you’re collecting the data. TEST the most important stuff first. Look at data you’ve TRACKED later on to optimize your campaign (keywords, time of day, etc)

Anyway, back to CPVLab.

Ultra Fast Stats

Looking at your Stats, Reports, and Trends in CPVLab is lighting quick. I really don’t know how the software does it so fast.

If you used 202 for CPV, then you’re probably familiar with the delays in reporting. Because there are so many ‘clicks’ with CPV, 202 took quite a while to display your reports if you changed any criteria. For example, if you had reports showing for the past week, and now you wanted to show the past 2 months, 202 had to chug along to run that new query. I don’t know if it was just me this happened to, but sometimes any new query would take several seconds, if not longer. It has been spruced up in recent versions, but its always something I noticed.

CPVLab produces your reports almost instantly. Change your time frame and your new report is showing as soon as you hit Go. Database optimization ninjas.

Excellent Support

Initially I, along with other users, had some problems with CPVLab. Since the software was “in-house,” then released to the public, there were a bunch of things that the creators didn’t notice. If you’ve ever built yourself some software, then you know your program like the back of your hand. But if you show it to some fresh eyes, they might go ‘wtf?’

Not only that, I have a knack for testing and breaking software. If your shit has a bug, I will probably find it. If your shit could be easier to use, I will notice it. If something is just wrong, I’ll say it.

Robert of CPVLab has taken my critiques, along with critiques from other users, to make CPVLab better. During my first few days of use, I would not have recommended this software just yet. It had many minor bugs and glitches.

For example, there was no time-frame on your stats. If you were on the fence and CPVLab, then bought it because of my review, then you’d be like “Fuck You Barman I can’t even view stats by day?!”

A recent patch a few days ago solved that problem and many others. There is another update in the works that solves a Reporting SubID’s problem (just use a pixel to get around this), along with a shit ton of other features I can’t remember.

I’m comfortable now in recommending CPVLab. There are a few issues still being worked on by their team, but CPVLab is ready to go out of the box.