Home > The Newbie Zone > Questions and Answers

---

Optimizing process (8)

---

Hello STMers.

Right now it's my first time optimizing a campaign. I was at -50% ROI, cut some landers (didn't have data to cut offers yet) but I'm at 0% ROI now. In what order should I cut variables like offers, landings, placements, banners, OSs, devices, carriers and similar when optimizing?

---

It sounds like the lander and offer you have now is working, if you're able to break even without cutting anything else.

I don't know how many offers you're testing right now, but if you have several running, I would suggest to wait until you have a winner.

After that, you can either cut your way to green, or test more offers. I personally would choose to test more offers first. But if you're breaking even, cutting stuff should get you into green easily, so you have that option.

But even now, you can drill down into OSs, device, carriers and see if there's anything doing horribly, that you can cut. Try to go easy on the cutting though, because you need traffic to test more offers and banners.

After testing more offers, you should be at the point where you'll be ready to scale. You can scale even now, but the better the offer is, the more sources you'll make profitable. It would be more wallet-friendly to use your break-even camp to test more offers first, find a better one, and THEN scale to other sources, as opposed to trying to scale a camp that's only at breakeven.

And of course you can do further testing on banners, landers, and the angles that drive them.

Without seeing detailed stats I can only give general advice. For more specific advice, please consider starting a follow-along.



Amy

---

Hello bnvltd,

First of all: good job! It’s great that you’ve been able to turn your first campaign from -50% ROI to a break even campaign!

It’d be great to get more info about your campaign. It’d help me make more suggestions. Even so, here are some tips.

When optimizing, you must gather enough statistics in order to make solid decisions. In order to make the right decisions, there’s something even more important than thinking about the order to cut: it’s looking at some key parameters, such as the CPA. Compare the CPA (of an offer, browser, device, etc.) with the payout and - after the campaign has been running for enough time to gather good statistics - you can follow this general rule: if you have CPA > 2 times the payout (or the cost, if it never converted) then it’s very hard to make it profitable and you should definitely cut it.

Besides CPA, CTR and CVR are also two key parameters. CTR measures the appeal of banners and landing pages, so you have to look at it to make the right decision. Moreover, you can’t forget that one banner with an overall bad performance may work very well with a specific LP, which means you must always analyse what’s the best combination banner+LP.

To help you making the decision concerning offers, you can also check the eCPM and CVR to see which offer performs best.

In addition, don’t think only about the cutting order but check first if there’s something horrible (as vortex said and as I also usually say) to cut, following the CPA > 2*payout rule. However, bear in mind that you have to understand how the cuts will affect other things and whether it will be better to cut a general thing, like a browser that’s used by several devices, or a specific device.

Remember: you’ve gotta let me know more details. For example, the number of banners/LP you’re using – a high number of banners and LPs can make them harder to optimize. The more banners/LP you use, the more you split your traffic among the various banners. Therefore, it takes more time to gather the info you need to make the right decisions.

More useful info: a) the type of offers you’re using; b) the GEO you’re working on; c) the spots you’re now attacking.

Hope I’ve been able to help!

Cheers!

---

Hello,

It's -10% actually now. And I played with my bid so I got no conversions now :/. I've used 5 landing, now I'm at 2. 3 offers and 8 banners. I'm promoting WA pins in South Africa

Thanks

---

Question: How many of all the different offer+lander+banner combinations are in green? How many of the green combos have 2 or more conversions?


Amy

---

Hello Amy,

Sorry for late answer. I have 8 profitable combos, only one with 2 conversions, rest with only one. I've got conversions until I changed a bid a little, and now I can't get back on track. Also after I cut 2 landers I've got only one conversion, but I cut landers based on statistical significant, so I don't think I've cut the wrong ones.

---

In that case I would suggest to keep running the camp. It's common to see a dive in conversion rates sometime after you first launch a camp - when you're getting traffic from so many placements, there's bound to be a lot of smaller placements that will get saturated very soon. There's always a chance you've "cut the wrong lander" because even with statistical methods nothing's ever 100% certain. I wouldn't recommend retesting landers. Actually, once you have a winning lander and 1-2 winning banners, I would suggest to test more offers if you can find similar offers. That is often the best way to potentially increase your ROI by large amounts.

Another possible reason for the lower conversion is you may have gotten outbid. Perhaps try increase your bid and see what happens.


Amy

---

I've added new offer and created five camps with different bids, so we'll see what will happen Thanks Amy

---

Home > The Newbie Zone > Questions and Answers