Home > Technical & Creative Skills > Tracking Campaigns

---

Problem with proper tracking because of cache (14)

---

Hi guys,

Okey, it's not a huge problem but it's definitely annoying.

Long story short I moved from affiliate network to direct advertiser.

I got a higher payout and I was like "okay, now I will make more money".

Just in case I did split-test. Affiliate Network vs Direct Advertiser. And after 1 day Aff network converted TWICE better!

I was like WTF. But I had a similar situation before. Back then I talked with AM and she said that aff network removed some filters and after she removed filters as well direct advertiser won.

And now I asked my AM the same thing. But she said that affiliate network uses exactly the same dashboard which I use. So there's no way they can remove some filters etc.

I did further investigation and that's what I noticed. When I click both links they look like these:

https://www.offer.com/afid=112&transid=398ea9673f

https://www.offer.com/afid=151&transid=398ea9673f

When I tried with different browser, no cache, then transid changed.

So my thinking is that people who clicked in the offer before, they have cache in the browser. This offer was running longer and that's why first offer from affiliate network worked better. Conversion goes to offer which was clicked earlier with the same cache.

Anyone experienced similar situation before?

I changed all traffic to direct advertiser because if it won't work, it's not worth running anyway.

But I had similar situation before, when because of cache final results were incorrect and wonder if that's common thing.

I know, I know it's not the most important thing in the world but I'm curious what about such split-testing think people who switched from aff network to direct advertizer.

Your thoughts?

---

I guess it´s because of the cookies.

It doesn´t matter if you run direct or through a network, ultimately the user will get redirected to the same offer.

When the user already has a cookie for that offer from a CPA network link and you send the same user then through the direct advertisers link to the same offer and the cookie from the network is still valid then the user would be connected to the network and not to the direct advertiser because that´s where the cookie was set.

I don´t have insights in the backend of such things but this is how I could imagine that´s how it is.

When you now run campaigns in small geos or very tight targeting and only on one source then there´s a higher chance that you target same users more often so that such situation can happen.

If I´m right with it then it would be a huge difference if for the advertiser the first cookie counts or the last one.

First cookie = CPA network still gets credit
Last cookie = The last set cookie, in this case the direct link, would get credit.

---

That's exactly the case. Currently I run on small targeting, 2 traffic sources etc. Thanks for confirmation that it could be the reason.

I will let you know if change all traffic to direct will help.

But in the same time it causes problems like:

1. I have campaign Sweeps 0.015
2. I duplicate on the next day and it's Sweeps 0.015 #2

and that's why 0.015 is always better for me and for some magic reason 0.015 #2 is worse.

So what we should look for is total profit for similar campaigns?

Like
all camps on bid 0.015 TS #1
all camps on bid 0.025 TS #1
all camps on bid 0.015 TS #2
all camps on bid 0.025 TS #2

etc.

---

I wouldn´t look for total profit.

I would rather see the offer from the CPA network and the offer from the advertiser as two different offers especially when they perform different.

When you just look for the total profit and the the direct offer isn´t converting good enough you risk that it eats the profits from the network offer.

Do you still run the network offer together with the direct offer?

Maybe you could ask the advertiser to basically copy/duplicate the offer for you on their platform so that it doesn´t rely on the same cookies as the original offer although I am not sure if it´s possible.

When it´s really cookie related you could also visit the offer url yourself and check how long the cookies are valid, could help to get a better idea.

Will try to brainstorm a bit more about it

---

Yeah, I would also do that. Or split test a fresh offer you have not run before (or same offer on a new geo / traffic source where you are less likely to have overlap). Because with the cookie, you have a bias toward the network.

Also, I would rename the thread to replace cache with cookies, so that it reflects more the discussion.

---

I wrote it too confusing I guess. In the part when I was talking about different campaigns etc. I was rather mean the same offer from the same network.

It was kind of different part of my post, let me explain.

When cookies (like @jeremie rightly said, not cache) stay the same (with this specific offer, whatever I run it with direct or CPA network), then sometimes I duplicate the same campaign, on the same traffic source, with the same settings. And duplicated campaign has lower performance than original one.

I was always sure that's how it is. But now I think that cookies could make that when in reality, both campaigns work similar.

Back to CPA vs direct.

Right now I just switched to direct advertiser offer. I gave it the whole day yesterday and it worked really well. Better than CPA network offer before.

Of course, I don't have 100% confidence when I didn't do normal A/B split test but it worked better than CPA network was working before so I will keep direct advertiser offer.

Besides, when CPA network use the same dashboard which I already use and they gives me lower payout + probably scrubs on their own it's small chance that I could make more with them.

I know, we always have to test but in this case real test was hard to do.

Next time I will ask about duplicating offer like you advised. But I remember that when I had similar problem in the past, advertiser couldn't really help me with similar cookies problem.

---

Do you run direct or with prelanders? If you run with prelanders, you should be able to set a cookie on your prelander and do your own tracking to see if you have situation where the same user passes by your landers from several different campaigns.

---

Yes, I run with prelanders. But cookie is set on offer page. Whenever I click on the offer link I see the same "transid=398ea9673f".

Is there a way to set it by myself with tracker and lander? If so, could you tell more about it?

---

They set their cookie. You could set yours for your own tracking purpose.

Your tracker shall also set a cookie to track unique/multiple visits. So you could have a special rule like in Voluum
https://doc.Voluum.com/en/rule_based...ue_visits.html

If you want to set your custom solution, you could do it in Javascript:
https://developer.mozilla.org/en-US/...ocument/cookie
https://stackoverflow.com/questions/...ith-javascript

Setting a cookie the first time, and checking if someone pass again by reading the cookie. When defining the cookie in Javascript, they usually expire after 7 days. If you want longer, you have to define them from the backend.

---

Hmm. Correct me if I'm wrong but I believe it works like this:

User land on LP -> Goes to offer and there's cookie set (let's say it's cookie transid=abc)

But he didn't convert for the first time. So process is similar when he lands again:

User land on LP -> Goes to offer and right now cookie is the same as it was before (transid=abc) -> This time he converts so advertiser sends postback for transid=abc.

So in the tracker I have counted conversion for 1st click, not 2nd.

Okey, I get that I can send user to unique LP when he lands again on the same LP. But it didn't change cookie on offer page, so anyway postback will be counted for 1st click (campaign). Because I can't change/clear cookie from user browser for advertiser's offer. And based on this cookie he sends data back to my tracker.

---

Yes, that´s how it works when first cookie counts.

When last cookie counts it would be like

User land on LP -> Goes to offer and there's cookie set (let's say it's cookie transid=abc)

But he didn't convert for the first time. So process is similar when he lands again:

User land on LP -> Goes to offer and new cookie overwrites the last cookie so that now cookie is transid=xyz -> This time he converts so advertiser sends postback for transid=xyz.

So in the tracker I have counted conversion for 2nd click, not 1st.

Yup, in the end you can´t do much about it because it´s not inside your funnel where you have control over.

It´s on the advertisers side so the advertiser would need to provide you with a "different" offer that uses a different cookie.

---

Yep, so I guess we have to live with that as advertisers can't really change these things. I asked them before and the problem is always "unfixable" for them.

Btw. From your experience, is there more offers with first or last cookies? Because from what I noticed, usually last cookie counts. Do you feel the same?

---

There are lots of methods actually. It is called attribution model:
https://support.google.com/google-ad.../6259715?hl=en
https://marketingland.com/what-is-at...odeling-254819

Each advertiser has his own policy, so it is difficult to say which is more common. It will vary between verticals / geo.

---

Good question, I never checked it for CPA offers.

But it would be fairer for the affiliates when it would be last cookie.

---

Home > Technical & Creative Skills > Tracking Campaigns