We’ve been using Facebook’s self serve platform for a few months now and has been making us a good amount of money consistently. One of the key things to our sucsess on Facebook has been testing, which is basically what affiliate marketing is.

We use a few tools, but the core of it is Prosper202 to get/read our data as well as some other tools and software we use, but we’re not going to go into that. So if you don’t have prosper202 already, go get it and install an SSL cert, it’ll save a lot of headache. Easiest way to get an SSL cert is just buy a domain with namecheap.com and get one free. Use the SSL pixels, it’ll save some headaches, but remember to update the subids daily as some will get missed!

So to start off, you’ll want to pick what your going to promote and start gathering some pictures for the ad, we usually do about 5 per test initially. As well as 5 pictures, we do 2-3 ad texts. So there will be a possible 15 variations per age group you target – not easy work as you gather, and it never is. If it’s too much for you, hire some one in-house to do it for you and pay them weekly or commision it. I’ve personally gone the commision route and it has worked well (more motivation).

Now the way to track this with prosper202 is to use their keywords. In the keyword you want to include a few things like the Age, Gender, Country, Picture version, Text Version. So format it something like this:

1. M-18-20-P1-T1-CA
2. F-18-20-P1-T1-CA
3. M-18-20-P2-T1-CA
4. F-18-20-P2-T1-CA

And so on, till you have a combination of everything. Keep in mind certain pictures and ad copies will make different demographics respond differently. For example picture 1 may respond very well to 18-20yr olds, but very differently to 30-35 year olds.

Once you have the ads submitted and approved, run a good sized test and make it significant so 1 click or 1 lead won’t throw off your numbers. So get some good statistical significance – it’s important.

Once you have the data, you want to go into Prosper and go to the Analyze tab and in the keyword box (in the More options section) put in M-18-20 so it only shows the KWs matching that. Then sort it by EPC (not ROI). If you keep prosper up to date, and update the CPCs the CPC data will set for example 50c to all KWs when in reality all the keywords may not be 50c cost. You will want to grab the combination with the highest EPC for each age / gender and run that. Once the winning one has exhausted, move on to the 2nd and so on.

The difference doing these tests can make is huge. To further these tests, you can start split testing landing pages also, as this will have a huge impact on the EPC too.

I can’t be bothered to write any more on this, so there it is! It’s been a great year for me and my company. I’ve met a lot of new people in the industry and learnt a lot this year. This industry is definately a great industry and the potential from it is HUGE.

When bidding, I generally like to get position number 1. On Adcenter position #1 converts FAR higher than position #3 for me, and has a brilliant click through rate. Don’t be afraid to pay $1-5 if you have to, just make sure it’s profitable. Quickly the price will drop. I bid $2, and pay half on average and it’s still going down.

Once you’ve got some conversions going through (with company name you’ll get 1 a day ATLEAST) then you ask for a bump. Easy as. Be good to your AMs, and they’ll be good to you. Luckily I’ve got great AMs (Minus Adam if he’s proven guilty). Check my Resources page to see what AMs I’ve got. Right now I’d highly suggest Josh Todd from CPA Empire – he is my favorite right now. Delivers all the time.

As for any campaign, tracking is key. I cannot recommend P202 enough. This stuff has delivered and saved me money numerous times. I’ve dropped expensive keywords that have not delivered once, which I probably wouldn’t of been able to do well or easily without P202. The stuff is free and secure, and usually updated every so often. The graphs can help get a quick overview if you want to see and go. The visitors tab lets you spy as well on visitors live. If your not using it, I suggest you try it. There are 2 versions, Hosted and Self hosted. I personally recommend self hosted and make sure you can keep it up 100% of the time.

If your not on AdCenter, use this £50 ($100ish) coupon to get started. The conversions with MSN are stellar. https://www.tryadcenternow.co.uk/ just fill it in and you’ll get emailed. I strongly suggest trying it. The support compared to AdWords will blow you away. I love their platform over YSM and Adwords. I generally start campaigns with MSN for testing (some lack of volume) then scale to YSM etc. MSN for me have had the highest conversions without doubt. On some days, 100% Conversion rate on some campaigns. So give them a go.

So your probably thinking, how is it done and how can I save, right? Well you want to group all the keywords into relevent groups. So heres a list of 10 keywords which should be put into 2 groups of 5:Visa Cards, Visa Debit, Visa prepaid, Visa premium, Visa lost, Master card, Master card euro, Master card stolen, Master card cost, Master card credit. As you can see it’s obvious that there are two Clear visible groups here. Master card and Visa. Along with these you want to make several ad variations and make sure the ads are all relevent to the group. For example:

Having the above for both campaigns isn’t going to convert as well as:

And you will certainly have a lower CTR and Quality score from Google. In YSM this means you’ll pay more per click too! So now that you’ve got two ads that will convert better, get better clicks you’ll get a better Quality score which will mean you can pay the minimum or a lower price for every click meaning you will get more for your daily budget. It also means that you will have a lower chance of being slapped by google. Because both groups would convert and perform better than one big irrelevant group, google’s automated system will be less likely to slap it. No slaps = cheaper clicks obviously.

So every now and then, you want to optimize and move keywords into smaller tighter fitting groups. Another way of doing this too is using multiple accounts and controlling it with a MCC account. You make several accounts, and then you would group them as “Good” “Poor” “Testing” or something in your head. When you start a test campaign, start them in your test accounts. If the campaign goes great, move it to your Good group, and make sure only well performing ads go into that account. If it only breaks even or makes a small loss, put it in your Poor account. This way Google will see your Good accounts an know your a performer and will slap that account less. It’s another way to keep your clicks cheap.

This is why you always need to optimize campaigns. It will boost your ROI% like crazy and keep you in the good books of Google. For YSM it will keep your clicks amazingly low and allow you to take advantage of the volume Yahoo hold. Not only that, it will keep your wallet heavier – which we all like. So next time you do keyword research, don’t just bung keywords into a group and remember to keep each group at 50 or under keywords.

]]> http://sohantanna.com/196/saving-with-ad-groups/feed/ 3