Those who dealt with networks for selling actions (leads), both on the part of the customer and the site, know that this is a thing in itself. In theory, everything is clear: there is an action, there is its price, bought-sold. But in practice, everything is much more complicated. Let's try to see the stones in this muddy water.
Why not work?
')
- large traffic and action losses
- instability and short period of the channel
- controversial accounting for multiple sources of traffic
- low conversion from traffic to leads
- long settlement between counterparties
- fraud, lead rejection
- a lot of manual work and mistakes
A little introductory text, then I will expand these theses.
CPA networks operation scheme
The network collects sites on the one hand and advertisers on the other, acts as an intermediary exchange between them.
1) The advertiser determines the action that will be the same action (the letter A in the abbreviation) and the subject of the sale.
2) Sites give traffic through the network to the advertiser.
3) Traffic is implemented in action.
4) Advertiser this action pays the network.
5) The network pays for the site.
This is very simplistic and superficial. There are a lot of types of traffic as well as types of actions - from a short request (like on a landing page) to filling out multi-step forms (loan application in MFI) and selling goods to an online store (all the steps of processing and payment).
Examples of large networks
Cityads.com
AdmitAd.com
ActionPay.ru
Several terms
Offer - lead advertiser application
Processing - the time of acceptance of the application after its receipt (20-40 days)
Hold - time of “settling” of the lead after its processing (20-30 days)
Postlink - the maximum time after which the lead can be counted from the moment a user transitions over an advertisement.
Diplink - link with an additional tag within the network.
Frode - fake, fake leads.
Yes, this environment is already overgrown with terminology, which further complicates the situation.
So, I expand my theses in order.
Loss of traffic and money
The main hole in the whole scheme lies here. There are many situations when the user makes an action, but the system will not count it:
- loss of the source of the transition (the link is not directly, through the site of the CPA network), the transition is technically not secured;
- the user followed the link, looked. After some time, I opened the site directly (through a search engine, bookmark, etc.);
- the user switched on advertising on mobile, walked around the site. He threw a link, opened it on a laptop and did an action;
- The user has exceeded the shelf life of cookies, disabled or erased them.
This is not a complete list.
Traffic loss:
- CPA network site is unavailable, freezes;
- counting script does not work correctly;
- the site gave traffic, and on the advertiser's side the site is not working (closed, updated, the promotion expired).
As a result, losses can reach 40% or more (you can write any number, the trick is that nobody knows it, including the networks themselves).
This leads to a loss of money (most often, for the site). And it goes on permanently all the time working with CPA networks.
An advertiser can say: And what difference does it make to me that there is lost if I pay for confirmed leads? Yes, but the whole scheme from this becomes less profitable and the site go. Who needs a stock exchange, where are some buyers?
Controversial issues regarding the attribution of the source to lead
Suppose the loss did not occur and the action correctly counted (the goods were bought, the loan was received). But in order to make it the user did not one click on advertising, but two or three (at different sites).
The debate about which source to refer to this lead is worthy of a debate between Plato and Socrates. You can read what Google Analitics thinks about this, to understand - to find the right is absolutely impossible.
What's happening? The network can (and, most likely, does just that) not take the lead to one and keep it.
Channel instability
The normal situation is when the advertiser set up the lead buying channel (offer), and he curled up not having lived for a month. Sites tried, saw the inefficiency and disconnected.
Or vice versa: the advertiser connected, did not fully understand how it works and withdrew his offer. All sites are forced to urgently redirect their advertising, look for an alternative.
That is, it is not necessary to expect that this channel will work for a long time and stably (as a context, for example).
Bad conversion
It should be noted that this item refers more to advertisers than to networks. But, we consider the whole scheme of work and the effectiveness of converting traffic to the lead is an extremely important point.
On average in hospital this conversion is bad. That is, the traffic goes, but there are no or little leads. As a result, for the site the cost of transitions becomes low and it turns off. It turns out useless work, distraction from other ways of monetization. Yes, networks will advise advertisers to help them become more efficient. But while the advertiser learns how to work well with traffic, either the donkey will die or the padish.
Long calculations
The scheme involves a multi-stage, each stage is a long time:
- The user must perform an action.
- The advertiser must confirm it.
- Hold to wait for no failure.
- Time from full confirmation to the order of payments.
- From the order of payments to the payment itself.
As a result, settlement with the site may occur in a
few months . This is the longest pattern of selling traffic in the universe!
Not to mention the fact that postpay itself in our realities suggests that you need to beat out your money.
How, in this case, the CPA network sells leads to the advertiser (in what parts, on a prepaid basis or in a small delay) an individual matter.
Some networks sometimes go for faster payouts, but this does not change the whole scheme. It turns out a constant, huge gap in money, for any business this is bad.
Frode, failure to take leads
There is a percentage of fraud on any network. That is, bids that are not created by users, but by unscrupulous market participants.
It was a noisy business when people (with names, phone numbers) who portrayed interested people were hired for a purchase of an apartment.
Fearing such situations, and sometimes just like that, advertisers are reinsured and discard suspicious applications. For this they have 101 reasons. Abandon leads make up 10-20 percent of the total number and are added to the losses in the first paragraph.
A lot of manual work and mistakes
It would seem that in this kind of automated systems everything should work by itself. Included advertising, she miscalculated and received money.
No, at each stage there is manual work from both the counterparties and the employees of the networks.
For example, in the
Actionpay network, in
order to confirm the applications, the manager needs to be written so that he meets with the advertiser (it is not known when), after that the applications (maybe) will be confirmed. Next, you need to wait for the withdrawal from the hold (check it in your account), how to withdraw - manually order the withdrawal of money. Then, you need to write to the manager again and he (when he wishes it is necessary) will make the translation. Natural brain removal.
All this is called the human factor, generating unjustified errors and time costs. The efficiency of the whole process is extremely low.
Summary
CPA networks today in Russia and the CIS really do not work. They are different, there is a difference, but this does not change the principle - most of the above points can be attributed to any. Advertisers more effectively build their schemes for receiving orders, without such intermediaries. And for sites to sell traffic using CPM and
CPC schemes.
Someone will say that I exaggerate and simply “do not know how to cook cats,” but I have the right to my opinion. What I use, actually.
PS I am pleased to read the counter-article from a representative of any network. But I strongly doubt the likelihood of such a public response.
PPS Intentionally did not write anything about the fact that the networks somehow cheated in their favor. Here the situation is as with cellular operators.