What stakes do you bid for Push Notification campaign?

[Win4ester]

Hello everybody!

A little story before the question.

Yesterday, I ran campaign of Push notification.

Offer = Sweepstake
Payment = 2,3 USD per lead
Geo = USA


I made some creatives, put the bid a little lower then reccommended, launched campaign. I deposited 100 USD on Propellerads. At first 15 minutes after campaign launched, nothing happened, I started to think, that I made a very low bid. Then, in a moment, when I refreshed page, I found out that I already spent 15 USD on campaign, what is crucial for me. At the end I recieved:

232 064 views
257 clicks
17.040 USD spent

1 conversion (2.3 USD).

And it all happened in like 15 minutes after the start of campaign...

So, the question is, what stakes do you bid while running Sweepstakes notification campaign?

 

[Graybeard]

You have to start the campaign paused.
Then set a day/time parameter -- every ad network has their own way of doing this or should ...
The best I have found so far is setting a max daily limit (low to start) and $x number of ads per hour.
Then start the ad campaign and make your observations.

Can you spend $1.00/hr to test first? (or some other low amount as a trial for a campaign)
Once you are confident $100/hr (or whatever)

(257/232064)*100=
00.1100%
My off-the-cuff observation:
there is no customer affinity --little interest to the ad
the offer itself has `mediocre` to poor conversion, or; the payout is too small.

If you redirect the traffic through you own server: then you have a complete log of who these 257 persons were; IP address, browser version --you could make more observations of the traffic qualities.

 

[Win4ester]

Can you spend $1.00/hr to test first?

I think, yes, but do you have any experience of bidding in this Push network? I want some advice about bidding, to be exact, if you could tell approximate bid, for spending 1 USD per hour. I know that it depends on CTR, and on the interest of audience to promoted offer, but maybe you have somekind of algorythm of testing Sweepstakes offers, for example, 10 creatives, and approximate bid for the first day?

 

[Win4ester]

2.3/257
.0089494163 = breakeven based on YOUR data

(2.3/257)*2
.0178988326 = 50% loss based on YOUR data
*added
keep in mind the low bidder gets the remnant if the traffic that is unsold at a higher bid. How is the quality sorted? Is it?

(2.3/257)/2
.0044747081= 50% ROI cash-on-cash based on YOUR data
Your data is empirical knowledge of the `KNOWN` results. Maybe, easier to think of this as historical results.

However, what do you know about the PEOPLE referred -- do you think that MIGHT matter in the equation? The *network* delivers *clicks*. Who are the clicks -- that matters in any algorithm.

There is no 'magic' number based on hypothesis only.
Predictions can only be made on theoretical assumptions or empirical data.

Thank you for the reply. No, traffic is not sorted, I failed trying to tune up the tracker to work properly, so, all I have - statistics of Propeller Ads and CPA network.

Thank you for calculations made for me, will use them in future. It was very surprising, when 15 USD ran away from me in one time.

 

[Graybeard]

2.3/257
.0089494163 = breakeven based on YOUR data

(2.3/257)*2
.0178988326 = 50% loss based on YOUR data
*added
keep in mind the low bidder gets the remnant if the traffic that is unsold at a higher bid. How is the quality sorted? Is it?

(2.3/257)/2
.0044747081= 50% ROI cash-on-cash based on YOUR data
Your data is empirical knowledge of the `KNOWN` results. Maybe, easier to think of this as historical results.

However, what do you know about the PEOPLE referred -- do you think that MIGHT matter in the equation? The *network* delivers *clicks*. Who are the clicks -- that matters in any algorithm.

There is no 'magic' number based on hypothesis only.
Predictions can only be made on theoretical assumptions or empirical data.

 

[Win4ester]

I don't know the exact interface at propeller.

However, I just made a CPC traffic bid at TrafficShop (mostly Adult) and I bid a new trial based upon the traffic and competition. As I raised the bid the numbers shown were updated in realtime (xrp AJAX) in the interface (I presume).
I think that is another way until you have hard data to work with.
500 -1500 clicks that you paid for is a statistically valid sample to work with.

Today, I finally found out, how to make tracker working, and surprised a bit. Why do I have only 10 clicks in campaign, and at the same time, tracker shows that there were 69 clicks and visits of the page?

 

[Graybeard]

I don't know the exact interface at propeller.

However, I just made a CPC traffic bid at TrafficShop (mostly Adult) and I bid a new trial based upon the traffic and competition. As I raised the bid the numbers shown were updated in realtime (xrp AJAX) in the interface (I presume).
I think that is another way until you have hard data to work with.
500 -1500 clicks that you paid for is a statistically valid sample to work with.

 

[Graybeard]

Cloaked bots (perhaps) Ask Me Anything - Testing Advertising Networks

post dealing with this subject >>> Ask Me Anything - Testing Advertising Networks