CASE: exclusive offers of Affmy in US with 55% ROI
Hey! Here we are with helpful information again! Today we are going to show you how to drive traffic from Mirelia - an ad network to our exclusive dating offers.
Traffic Source: Mirelia
Advertising format: banner ads
Network: Affmy
Offers: JoinTheDating, MilfNearMe, LookAWoman
Cost: $359.01
Revenue: $559.8
Profit: $200.79.
ROI: 55.93%
How to choose a traffic source
This is our first case of driving traffic from banner ads. To test we chose Mirelia, and there are several reasons for it:
- high quality traffic from direct sources
- a personal account manager
- a possibility of API integration.
- its own pre-landing pages for split testing.
- a possibility to buy traffic from different segments, sorted by price.
How to set a tracker
We use Keitaro tracker to monitor statistics and optimize campaigns (this is necessary in the future) First we add a source: go to source tab and click on “Create”.
Further we fill in “Name” section and add parameters. We have to enter them manually in the link, since the ad network does not have macros, so we will create a template to make it easier to fill in data in the future. Along with the standard parameters, we add the Preland ID (we will write the ID of the pre-land that we use in the ad campaign (AC) and the Segment (for the indication of the traffic segment and the convenience of analytics).
Note. Traffic loss as a default is 0%. This parameter reflects a percentage of traffic loss and it can vary in all ad networks. The thing is there can be too many or too few clicks in the networks, that’s why actual costs may differ a bit from the costs shown in the tracker. It is better to update costs in the tracker as they are, for example weekly or monthly.
To do it, click on three dots near “Campaign” tab, and then click on “Update costs”.
How to set a campaign
We start to set the ad campaign in Mirelia with selecting a pay type: CPC (cost per clicks) or CPM (cost per impressions). You will find some distinctive features of their launch.
If you use CPC, please note:
- to choose a segment where you are going to drive traffic (the segments are divided according to the cost);
- to choose pre-landing pages for split testing;
- to set CPC.
If you use CPM, please note:
- to choose a location for a banner - Top, Interstitial and Bottom;
- to upload creatives (no more than 30 pieces);
- to set Bidding priority. The higher your bid, the more traffic you get.
We have consulted with the managers and they recommended starting with CPC, as this pay type has simpler settings, there is no need to add creatives and to monitor which one is more convertible.
To create a campaign you should first of all choose a GEO (in our case it’s the USA) and to name the ad campaign. For easy stats monitoring in the ad network, we call Mirelia_US_cpc_brown1, and edit the last part in accordance with AC settings (segment and ID). We test only mobile traffic, as most of the audience uses their cell phones. Choose Dating as a Vertical.
Then we copy the campaign link and paste it in Mirelia. You should add the parameters, which you have chosen in the settings: cost per click, ID, pre-landing page, segment and campaign number.
To get a campaign number, first you need to save your AC and then to look at its number in «Campaigns» tab (see the screenshot below in the right).
Then we go to traffic segment choice. In Mirelia we are offered 9 user groups and traffic from the back button (a button used to return to the previous page), they are situated in the ascending order by price in the list. In the first tests we are going to try the first 3 groups: Brown, Grey and Green, launching them with the minimum CPC to analyze the source.
After source testing you can increase budgets and test more expensive segments, we do not recommend starting with expensive traffic.
The next settings section is connected with limits. Everything, underlined with red color on the screenshot, must be filled in, the remaining is optional. Then we set the minimum CPC for the segments: Brown – $0.01, Grey – $0.02, Green – $0.04. It is better to test with minimum bid not to overdose on traffic amount and not to run at a deficit.
The total budget in money is not limited: it is calculated as cap by clicks in Mirelia. We set 2000 clicks per each campaign, and it will be clear after the optimization, whether there is a need to change this data. We have not set a campaign daily or hourly cap yet because we need to know what traffic amount we are going to buy in 24 hours according to the set CR.
Further we need to fill in the “Domain” field - it is a unique name of the campaign group (and not a domain name) in Mirelia. We write any name that we haven’t used before as a domain, for example, dat.com (each company should have its own domain). It is necessary not to buy traffic twice. We do not set domain frequency cap, campaign frequency cap only.
We set a campaign frequency cap per 1 user only once, it is an optimal indicator for source tests.
The next step is to set time targeting. We set the time zone as UTC+03:00, as the tracker is set according to the same time zone. The campaign runs 24 hours 7 days a week during the test, and if after the test drive we see that the AC is more efficient in a certain period, we set targeting by certain days or hours (but according to our experience, to do it, we should collect a lot of statistics).
Then we chose a pre-landing page. We recommend testing 5 different pre-landing for 1 traffic segment to define the best and to drive traffic to it. It is necessary to distribute traffic equally in A/B split testing: we select 1 pre-landing page and 1 digit in «Traffic distribution» settings for each campaign. The first AC has 1, the second - 2, etc. up to the fifth. Such a distribution means that each pre-landing page has ⅕ of the traffic, and the statistics is more illustrative.
We talked to the managers, and they gave us the list of the 14 top pre-landing pages, we chose the first 5 of them to test the traffic. The ID of the pre-landing pages we have chosen: 37, 34, 18, 57 and 36.
So we have 15 campaigns - 3 traffic segments, each of them have 5 AC with various pre-landing pages.
Now we go back to campaign total cap: we have set 2000 clicks per campaign, in total we have 15 campaigns, it means we set 30000 as the maximum amount of the clicks. We also realize that there can be fewer of them. We have 3 traffic segments, it means that 1 segment has 10000 clicks.
The first statistics summary and optimization
We have got the first period of stats monitoring in 2 days after the campaigns launch, as there is enough data for optimization by this period. Statistics by offers showed that the minimum ROI got LookAWoman – it had a negative result. The other two offers pulled away a lot, so we leave only them, giving 50% of traffic to each.
The «Green» segment had the highest ROI - 27, 1%, but the largest amount of traffic had the «Brown» one. It can be connected with the fact that Green is of higher quality, that’s why there is more rivalry, and we bought not so much traffic at the minimum bid. But for more accurate conclusions, we need to collect more statistics.
We also monitored the general stats by pre-landing pages. We have 2 pre-landing pages with the negative ROI, switch them off and keep testing the remaining three to define the leader.
We also had data sorted by segments and preland IDs: two first places had 34 and 57 preland IDs in «Brown» segment, that’s why we distribute ⅖ of traffic to each, and to 37 preland ID - ⅕ .
The same situation is in «Green» segment:
we drive ⅖ of traffic to 34, ⅖ of traffic to 37 and ⅕ to 57.
In “Grey” segment:
we drive ⅖ of traffic to 34, ⅖ of traffic to 57 and ⅕ to 37.
Note: the campaign settings are situated in «A/B Split test» section, in «Traffic distribution» field.
The second statistics summary and optimization
After 4 days we have had the second statistics summary and optimized once again the campaigns, since we have enough data. As a result we have the following statistics, sorted by offers:
JoinTheDating got 49, 39% ROI,
MilfNearMe got 27, 56% ROI.
We are going to drive some more, making another optimization by prelands to make sure that JoinTheDating is a true leader.
There are no changes in stats, sorted by segments: «Green» got the highest ROI - 35, 98% again.
We have the leader of prelands - 37 preland with 86, 99% ROI.
That’s why we leave only one campaign with one pre-landing page in each segment and in the "Traffic distribution" field choose numbers from 1 to 5, so that all traffic can go to campaigns with these prelands.
We are satisfied with the traffic quality and planning to drive it further, that’s why we increase campaign total cap up to 5000 for each campaign, which remained.
The third summary of statistics
Since 12 days after we have driven traffic to the campaigns, we got the following results:
Cost: $359.01
Profit: $200.79
ROI: 55.93%
JoinTheDating got the highest ROI – 62, 51%, it means that it is a true leader, and we are going to use it further as the main offer.
According to traffic segments, we got the following results:
“Green” segment got 62,19% ROI;
“Brown” segment got 58,08% ROI;
“Grey” segment got 46,11% ROI;
As we have launched the campaigns with the minimum CR, we can guess that the more expensive traffic, the higher its CR. The source test has shown that even cheap segments give positive results, so you can launch campaigns with more expensive segments.
As a result of traffic drive, the top preland got 74, 03% ROI. Further you can test new ones, if this preland has worse conversion.
