1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Nail & Scale - The Art of Blowing Up Campaigns With Big Data

Discussion in 'Making Money' started by MrE, Aug 25, 2015.

  1. MrE

    MrE Executive VIP

    Joined:
    Sep 16, 2008
    Messages:
    838
    Likes Received:
    2,770
    Nail & Scale - The Art of Blowing Up Campaigns With Big Data

    This is going to be a long one, this is written as an example for the BassTrackerBoats FB group posting thread to illustrate but can be applied to any type of marketing to capture data, then nail & scale. Skip this if you're not currently making money with this method or not running any type of campaigns generating revenue - advanced material to follow. If you're not running FB group posting then skip to prosper202 section below.

    I just added a new FB groups campaign the other day so I wanted to share my process for cutting straight through the bullshit and nailing it right away. This thread will discuss some more advanced strategies, data collection methods, how to analyze/use that data, some small case studies in LP optimization, and the straight line for what everyone cares about.... a winning campaign that blows up AND can SCALE. This is material that will blow your shit up if you're already earning with this method. What this will not teach you is offer or group selection processes, honing your eye for that is something that comes only with practice.

    Everything in IM is information & data, the more relevant and accurate data (feedback) you have the tighter you can dial things in. That means less wasted time, less frustration, less wondering why something isn't working, and a map straight to the pot of gold. You can immediately identify the point of failure in your campaign by analyzing the data - bad lead ins? bad group selection? poor LP? Offer sucks?

    Let's get started, fast forward and I have:
    - selected a new niche I believe has a lot of potential and got a PPS/PPL (CC) offer with a high payout
    - selected an offer which solves a problem and offers value to the customers
    - checked the LP/offer on mobile devices (bulk of FB traffic) to ensure no compatability issues (http://www.mobiletest.me for devices you don't have handy)
    - joined 40 groups in related niche for a small scale test
    - setup my news/review site LP while waiting for group acceptances
    - wrote 10 unique lead ins for my test
    - installed analytics on the news/review LP site
    - setup prosper202 (free self hosted ad/click server webapp) on a different domain

    Take a few mins and create a short custom tracking code for EACH GROUP, doesn't matter how you do it, you can use the last few digits of the group, make a custom naming convention you have in a spreadsheet, etc. I.E. group number is 398593942093202 your unique code could be 93202, or 202-3 (3 indicates which account your posting with to this group). Assign each lead in comment to a group ID and toss in a spreadsheet, you should now have 40 full group links, short unique code for each, a comment beside each in a spreadsheet, a note of which account will be used and the time posted. When posting your comments drop your URL with the unique ID added on the end for each corresponding group. I.E. http://www.mysite.com/my-miracle-article/?click=93202 - every post you make has a different tracking id on it linked to the FB group.

    When traffic starts rolling in you can now see which groups are delivering traffic and which are duds, ok that's pretty useful.... so what's the big deal?

    [​IMG]

    NEXT, you're going to use php and $_GET['click'] or whatever your ?variable= was named to pass the FB unique group ID to your LP. Now setup a unique name for EVERY TEXT LINK & EVERY BANNER/IMAGE LINK in your LP, i.e. text-link1, text-link2, adchoices-body, video-sidebar, sort out unique names for each link and have a clear record. NEXT.

    You're going to setup a new campaign in prosper202, fill in affil network name, campaign name, payout, grab affil link from your network with the SUBID function on the end, this is important. (Google how to setup a new campaign in prosper202 if you're not sure or lost)

    Next, go to #8 Tracking Links within prosper202 dashboard and grab your tracking link with the C1 variable filled in, if done correctly your link will look like: http://www.yourdomain.com/tracking202/redirect/dl.php?t202id=xxx&c1=&t202kw= - perfect let's move on.

    Now in your LP you're going to drop this link, after c1= put your link name on page, text-link1, adchoices-body, ad-sidebar, blah blah. and then after kw= put <?php echo $click ?> - this is passing the FB group ID & which button they clicked on your site to the adserver. Google if you're lost or hire a freelancer to help you with technical snags. PHP Code plugin is good for using php within wordpress posts/pages/widgets. You should have many different links on the page now corresponding to the link type and location so you can see exactly what was clicked.

    NOW FOR THE FUN PART:

    Test and make sure all the links are setup properly on your LP, click through and make sure you end up on the affiliate offer page. Now take a look in prosper to see that the data was properly recorded, also check with the affiliate network that the SUBID field is passed on to them and you can see it in their stats. If everything looks good start posting your FB ads with unique tracking ID on the end.

    Once traffic starts rolling in you can see group data in analytics, but prosper202 data is the real gem. You will see time of click, device type/browser, location, IP, the referring URL, which link they clicked on your LP, which FB group they came from, which lead in text you posted with, which account was used and what time the post was made. All of this is recorded on your prosper202 ad server and assigned a unique SUBID which is passed to the affiliate network (while cloaking your original LP), you can now see EXACTLY WHICH USERS CONVERT.

    [​IMG]


    Fast forward to the end of day 1 on my new test campaign, posted to 40 groups (the more the better for accurate sample size) and now I'm reviewing my captured data. You can see which subIDs converted in the affiliate network stats, and you have linked campaign data for each user in your prosper202 backend... let's take a look.

    Unique hits to my LP: 613
    Unique clicks to my offer page: 221
    CTR: 36%
    Leads: 7
    Rev: $200+
    Tablet/Desktop EPC: $1.70+
    iPhone EPC: $0.73
    Android/other devices EPC: $0.00
    International traffic (Non US/CAN) EPC: $0.00

    [​IMG]

    I can see that out of 40 groups, 33 produced clicks, while only 4 groups resulted in leads. (One group crushing it with 4 leads) This doesn't mean I'm writing off the groups that didn't produce leads, but at this point I weed out any groups that didn't result in a single click. I keep 1+ clicks for now for further testing, after a week or two the poor performers are cut. I consider this campaign an immediate home run based on the $5+/per post revenue target being smashed, and immediately seeing clear lead ins/type of groups that work well, seeing which elements of my LP hit and missed and knowing the offer converted well for two segments of my traffic.

    [​IMG]

    At this point you can dial it in and see what type of comments/groups are working best and dig deeper to figure out why. You can also try changing up non performing links/creatives on your LP, and split testing to see if you can juice CTR, EPC, comparing new data to your previous data for all changes. The more data you gather & analyze the better, it is your map directly to the pot of gold, no detours.

    After you have thousands of clicks to your offer the data you have starts to point to clear truths, and you can seek new monetization paths for poorly performing sub-segments. Example: if android devices continued to perform very poor over a large sample size then I would try serving adsense ads to Android devices. If International traffic didn't convert at all then I could try adsense or re-direct them to a LP for a different product in the same niche. Micro optimizing for the poorly performing segments and keeping the winners running strong, that's the key to squeezing out every last penny from your traffic and having SKY HIGH EPC's.

    PART II will focus on the scaling aspect of the FB campaign and PART III will be split testing results from my campaigns which will help juice your revenue. But right now.... GET TO WORK.


    TLDR; IM IS DEAD AND THERE'S NO MORE MONEY TO BE MADE, APPLY AT MCDONALDS FOR REAL CASH.
     
    • Thanks Thanks x 112
  2. soccerlover

    soccerlover Jr. VIP Jr. VIP

    Joined:
    Jun 12, 2014
    Messages:
    3,181
    Likes Received:
    1,682
    Gender:
    Male
    Occupation:
    Seo Analyst :D
    Location:
    ♥♥♥ BHW ♥♥♥
    Home Page:
    Great and Detail perspective.
    Nice Method.
     
  3. secretboy08

    secretboy08 Jr. VIP Jr. VIP

    Joined:
    Apr 2, 2008
    Messages:
    2,393
    Likes Received:
    881
    Occupation:
    Internet marketer
    Location:
    Global Citizen
    Wonderful. Great tutorial. I do not follow the fb method, but it is useful for my normal sites, atleast the landing page to offer tracking. :)
     
  4. MrE

    MrE Executive VIP

    Joined:
    Sep 16, 2008
    Messages:
    838
    Likes Received:
    2,770
    Exactly, this can be applied to any type of marketing campaign to capture more data - mo data mo money.

    Anyone familiar with mediabuys, PPC, or PPV campaigns should have some experience with a similar type of ad server setup and knows how vital it is to success.
     
    • Thanks Thanks x 2
  5. RuthSam

    RuthSam Jr. VIP Jr. VIP Premium Member

    Joined:
    Mar 19, 2010
    Messages:
    3,812
    Likes Received:
    973
    Gender:
    Male
    Home Page:
    Nice share, thanks mate :)
     
  6. tompots

    tompots Elite Member Premium Member

    Joined:
    Dec 11, 2011
    Messages:
    4,371
    Likes Received:
    3,964
    Gender:
    Male
    Occupation:
    Full Time Bot Developer
    Location:
    Automation Alternatives
    Home Page:
  7. wealth54

    wealth54 Regular Member

    Joined:
    Jan 7, 2009
    Messages:
    314
    Likes Received:
    100
    Occupation:
    IM
    Location:
    PC
    awesome share MrE ! thank you :)
     
  8. Ambitious12

    Ambitious12 Elite Member

    Joined:
    Jun 26, 2014
    Messages:
    3,097
    Likes Received:
    608
    Occupation:
    No Occupation
    Location:
    Among the Stars
    Thats really great method I would say.:)
     
  9. Boriss

    Boriss Supreme Member

    Joined:
    Nov 7, 2009
    Messages:
    1,427
    Likes Received:
    562
    Location:
    Inside a Monitor
    Very detailed post MrE. Love it.

    Thank you for sharing this knowledge with us.
     
  10. MrE

    MrE Executive VIP

    Joined:
    Sep 16, 2008
    Messages:
    838
    Likes Received:
    2,770
    PART II - The Power of Scale

    After you have seen some positive and encouraging results from initial testing it's time to scale up and hit some big numbers. This post will for the most part apply to the BTB FB Groups method mentioned in the first post.

    Step 1 - The 3 Minute Profile Optimization for Increasing Group Acceptance Rates

    - Hide your FB friend's, (show only to friends or yourself)
    - Like a few things for each type of interest and then set all to public (books, tv, movies, sports, interests)
    - Like some topics related to your niche
    - Go to pages feed and share 5 or 6 posts to your wall (set as public)
    - Go to your wall - from drop down arrow on right side of each post EDIT DATE - spread them out going back a few months

    [​IMG]

    The edit date can still show the original date but for a quick group admin looking at your page it work's wonders for giving an aged real account appearance in minutes, this will put your group acceptance rates way up. Start putting out requests to join your groups and wait for acceptance, once you're accepted to a few groups go to your profile and set groups to show only friends or custom. Most real accounts aren't in hundreds of groups and you'll likely be viewed as a spammer dropping future acceptance rates.

    Step 2 - Follow the Rare White Rabbit to the Promise Land

    Everyone knows how to scrape groups through the search, or using automated tools and this is good for returning the low hanging fruit groups with obvious keywords. Once you've been accepted in a bunch of groups start opening other peoples profiles who have posted on the group pages, many people don't block groups they're in and when you hit the "rare white rabbit" they deliver the jackpot on niche groups you would never find by scraping. These groups can be local geo areas, not have any keywords you've searched or be loosely related sub-niches that you didn't even think of; the important part is they showed you possible common interest groups.

    If you can't find these groups by scraping, it means others won't either and that means less spam on the page and higher user engagement. If pages are getting slammed with posts frequently many users will disable group notifications, so in many cases the BIG groups are the worst performers for your campaigns. What works best is smaller groups with very low spam, and very high user engagement. I'll take 1000 - 2000 eyeballs on my post in their newsfeed over a large group any day of the week, and crunched numbers support this.

    [​IMG]

    You can hire a VA or create a quick macro script to scrape all groups once you hit on good accounts, this part isn't fun but this separates the men from the boys, the motivated from the lazy, and the winners from the losers.

    Step 3 - The Trojan Horse & Useful Idiots Approach

    After you have the test data from the original post relating to conversions you can see which type of lead in comments work the best to prime the user to buy. Scrap all your other lead ins except for the best 2 or 3 performing ones, now you're going to re-write variations based on these until you have 15-20 very similar lead ins which are high converters.

    At the end of each lead in you're now going to write one more sentence which demands user action, in this case some type of interaction (Like/Comment/Share) with your post. The reasons for this are:

    - Spread the reach of your post to their friends (social proof coming from a friend)
    - Amplifying the reach and exposure within the group (your post will get more visibility and stick up top longer the more interactions)
    - Safeguarding your FB accounts and building rep (What spammers get hundreds of likes/comments/shares on their spam?)

    [​IMG]

    In my experience my accounts haven't encountered any issues once my posts started focusing on these types of interactions, hundreds of likes/shares/comments will offset the odd few spam complaints you get IMO. Anyone else have thoughts on this? My thinking is if you're getting booted from groups for spamming or getting reported that you're A) not targeting groups properly B) being too obvious/aggressive about your lead in or C) have a scammy looking offer/page - I would re-think this part if you're not being received warmly.

    TELL THE USER EXACTLY WHAT TO DO - Examples: "LIKE & SHARE IF YOU'RE MOTIVATED TO IMPROVE YOURSELF", "LIKE & SHARE IF YOU'RE A HARD WORKING SINGLE MOM", "LIKE & COMMENT 'STOP HUNTING INNOCENT ANIMALS' IF YOU LOVE ANIMALS", frame the writing in a way that users will have a hard time disagreeing with, and almost make them feel like a bad person if they don't agree with your obvious statement. In some niches that aren't professional in nature adding in a few FB emoticons helps, and being the first to comment the phrase on your post can start the freight train rolling down the tracks.

    Let's Recap:
    - "age" legit looking new accounts in minutes
    - find the hidden gem groups and get quick acceptance
    - spread your message further than just 1 post with the friends "trojan horse shares" and magnify reach on groups by utilizing "useful idiots" to further your cause

    All that's left is getting more proxies, buying more FB accounts (Good Provider Here), automating with FB Dominator or MP and putting in the work to scale.

    Constantly keep reviewing your captured data and split testing to optimize your LP, and keep increasing your traffic levels to make it rain. It's not complicated, just put in the time.

    PART III coming soon will discuss real world case studies in juicing CTR's, EPCs, & tips on squeezing every last dime out of your campaigns from a few different angles.




    TLDR; IM IS SATURATED AND NO ONE IS MAKING MONEY.
     
    • Thanks Thanks x 42
    Last edited: Aug 27, 2015
  11. soulcollector

    soulcollector Senior Member

    Joined:
    May 10, 2014
    Messages:
    833
    Likes Received:
    301
    Location:
    USA
    I normally don't use CPA offers when doing this, and I get great results when promoting eCommerce sites with something similar to this.
     
  12. secretboy08

    secretboy08 Jr. VIP Jr. VIP

    Joined:
    Apr 2, 2008
    Messages:
    2,393
    Likes Received:
    881
    Occupation:
    Internet marketer
    Location:
    Global Citizen
    I just tried out the Rare White Rabbit approach and it sure gave me some nice targets which had skipped the scraping.

    Another great gem of a post from you. :)
     
  13. THUNDERELVI

    THUNDERELVI Elite Member

    Joined:
    Sep 12, 2009
    Messages:
    2,399
    Likes Received:
    2,074
    Gender:
    Male
    Location:
    W3
    Subscribed to this thread for part 3 :D Part 1 is real gold, nailing down on winning campaigns is the key to spam exactly where is needed!
     
  14. CoolAdvisor

    CoolAdvisor Senior Member

    Joined:
    Mar 24, 2008
    Messages:
    1,010
    Likes Received:
    371
    This is must read and will work with other social networks too.
     
  15. ProNe55

    ProNe55 Regular Member

    Joined:
    Aug 6, 2011
    Messages:
    207
    Likes Received:
    65
    This is a really good post, thanks for the share! have bookmarked it.
     
  16. Freddyjames

    Freddyjames Senior Member

    Joined:
    Jul 4, 2011
    Messages:
    842
    Likes Received:
    1,546
    Just. Fucking. Awesome.

    Thank you.

    I really hope that this method is as long-term as I believe it will be. i.e. limitless.
     
  17. unitedforever

    unitedforever Jr. VIP Jr. VIP

    Joined:
    Jul 11, 2015
    Messages:
    125
    Likes Received:
    120
    Gender:
    Male
    Location:
    United Kingdom
    Thanks a lot for this. I'm going to put this into practice :D
     
  18. GreyHatNewb

    GreyHatNewb Junior Member

    Joined:
    Nov 26, 2014
    Messages:
    144
    Likes Received:
    27
    More valuable gold for us newbies. It would be very hard for any newb to not succeed with info like this
    #Subscribed
     
  19. Paper-Boy

    Paper-Boy Elite Member

    Joined:
    Jun 17, 2009
    Messages:
    5,118
    Likes Received:
    1,823
    Seriously bro? This post is the epitome of spoon-feeding to a toddler, haha. Probably one of the best posts I've seen in awhile.

    You can use this info and apply on so many other things.

    Thrive is also a good self-hosted tracker. Been using it for awhile now and so far so good.

    If you don't care too much about your data and want a simple, fast, reliable tracker, then voluum is the way to go.
     
    • Thanks Thanks x 3
  20. MrE

    MrE Executive VIP

    Joined:
    Sep 16, 2008
    Messages:
    838
    Likes Received:
    2,770
    @Paper-Boy, Thrive looks really good but haven't played with it yet myself. This type of example is vague enough that a lot of information can be revealed that is broadly applicable without outing my campaigns or hurting income.

    I'm a big fan of self hosted data as this industry is dirty as hell and you're very naive if you think letting someone else have your exact campaign data is a smart idea so I'd stick with prosper or thrive.
     
    • Thanks Thanks x 2