Case: What to do to reduce the percentage of trash and increase the number of leads to earn $ 67,000+

7 min readFeb 12, 2020


Offer: Erogan

Bundle: Pre-landings and landing pages from an affiliate program that we have technically upgraded

CPA Network: Leadbit

Traffic Source: Facebook (Target)

Drain period: 12/01/2019–01/31/2020

GEO: Italy, Romania, Spain, Portugal, Czech Republic

Spent: $ 24100 (~ 1.5cc rubles)

Received: $ 67767

Profit: $ 43,667

ROI: 181%

Technical Part

Let us quickly remember once again what we use in our work if someone has not read our previous cases:

  • AEZAKMI antidetect browser
  • Mobile RU-proxies (own farm);
  • Facebook accounts from AEZAKMI;
  • Cloaca;
  • Creatives from NEOCREO (channel in TG:;


NPPR.TEAM is still here and we are going to tell you what we did during the period from 12/01/2019 to 01/31/2020. And we did a lot, so we did not get in touch at the end of December. This applies more to the refinement of partnership pre-landings and the development of our own technical solutions for working with traffic.

1) The first thing that we did, we finalized all the landings. Now, there are application forms on ALL of our pre-landings, that have the common form according to our tests. It turned out to bring out universal combinations, which are automatically substituted for our pre-landings, i.e. the forms themselves are dynamic and can be different types — on one pre-landing there is a vertical form with one type of button, on the other horizontal, on the third mini-quiz, and so on.

We have added automatic region prefixes on all forms. The user needs to enter even less information and this increases the level of validity of phones.

We removed the text keyboard on mobile versions of our pre-landings in order that all users would automatically see a digital version of the keyboard when entering a phone number. As a result, we improved the convenience of entering the phone and reduced the number of trash, because sometimes people miss out on small buttons, and sometimes even enter some kind of comments in brackets in the phone field.

2) The second thing that we decided to do was to additionally warm up all the leads using SMS. Actually, the decision to add prefixes to forms was associated with this item, since without this prefix, there were a large amount of invalid phone numbers when sending SMS.

Now all the phones that are left through the forms are tested for validity and if the phone is valid, then a warm-up SMS is automatically sent to it.

To send an SMS, we used API of the Send Pulse mailing service.

3) The third change is more global. We have implemented new functionality in our working panel. This change is an automatic traffic distributor between our pe-landings.

When we were analyzing traffic, at a certain stage we were faced with the problem of repeated visits due to which there were drawdowns. For your understanding, some users, according to the metrics, have visited our pre-landings 10–20–30 times and even in spite of it they finally left valid contact information.

We decided to make the most common marketing trick is to make a large number of different touches with one user, so that with a high probability to get a lead.

Now when a new user logs into any of our pre-landings, we saved his cookies and based on these data, the automatic distribution of traffic between our pre-landings is enabled. Now, on average, we have raised 8 pre-landings per GEO, between which distribution occurs.

In other words, when a user visit us and our system sees that he was already on our first pre-landing, it shows him the pre-landing number 2. If the user does not convert from the second time, then the next time he will receive the pre-landing number 3. And so on.

This approach allows you not to be afraid of repeated visits and convert any traffic both new and old.

These are the technical changes we have implemented in these 2 months. Now let’s move on to how it was used.

How we did it

The approach to traffic arbitrage have noticeably changed, since now we have to use accounts with a $50 limit, but with a large number of accounts.

Developers of the anti-detection browser Aezakmi helped us with this situation. They prepared and released API for Aezakmi. We were invited as beta-testers of the new functionality. We did not refuse, because this ultimately accelerated our work very well.

What did we manage to automate with the new Aezakmi functionality? Now we automatically generate the fingerprint profiles with cookies of the accounts.

Now we do not bother with manual preparation of the launch, but simply generate 100–200–300 fingerprint profiles and start.

The only thing that upset us a little bit was that Aezakmi Facebook accounts with BM was out of stock. But the guys gave us accounts without BM, also of their own production, for tests. After we conducted our tests, we realized that we were upset in vain. These accounts can work just as well. Therefore, if someone doubts about these accounts, then try it by yourself. But in any case, the guys have already begun the first tests of new accounts with BM and promise to soon restart the sale of accounts with BM.

Accounts now live less than in previous periods. Now, if an account in our realities lasts 3 days, then we believe that this is already good.

At the same time, the price of leads does not increase, which is great. The average price of a lead during this period was around $3–4. On screens you may see that during this period, all buyers from our team successfully chilled out, so we suggest watching cartoons:

We used almost the same creatives as in our past cases. We won’t show it again here. Who cares, see the creatives here and here. The general approach to creatives will be clear for you.

New GEO Testing

In January, we decided to test new GEOs for us — Spain and Portugal. And also continued to work with the Czech Republic.

For tests, we simply moved our current approaches from Italy and Romania, which gave different results for these two GEOs. Spain proved to be very good in terms of the price of lead in terms and in terms of approvement. Portugal is the opposite. Yes, there was no loses, but the profit on this GEO was minimal.

Spain and Portugal Test Results in the Last Week of January

We take Spain to regular work, as it has proved to be good on the available approaches, but Portugal continues to test and look for approaches that will give more adequate indicators.

General screen of stats from CPA Network:

The average approvement in Romania was 61%, in Italy — 48%, in Spain — 53%, in the Czech Republic — 44%, and in Portugal, which was unsuccessful for us, — 29%. There were no questions regarding the work of the Call-center during this period, except for one case — someone increased all the limits in Italy and we suffered a couple of days, since the Call-center simply could not manage it.


The general conclusion is that it has become more difficult to work, but thanks to a systematic approach, we managed not only to get loses, but also to increase indicators.

Now we are actively working on the technical part and I think that we will soon talk about the next major update.

Also, since the new month, we have been taking new offers according to the SS (Straight Sell) model, where people do not leave a request, but pay on the site. Therefore, soon wait for comparisons of offers from the same vertical but with different models — COD and SS.

P.S. In this case, we wanted to show our work with pre-landings and the technical part, so do not swear at such a reflection of creatives and screenshots from Facebook.

Link to our group:

See you!