Home > Paid Traffic Sources > Native

---

My $5K Outbrain test (and failure) (16)

---

So during the past weeks I've spent (wasted) about $5,300 on Outbrain and no results.

I'd like to share it with you guys, so hopefully, you'll be able to help us out.

Few things to keep in mind:

1. Offer works very well (on FB, IG, Liveintent, Google, Linkedin, and even managed to break even on Quora with small $450/day budget).

2. It's an Ecom offer

Here's the funnel:

AD 》 advert 》 offer page》 Checkout

Quite basic...

Here's the setup:

1. We installed pixel for offer page visit, checkout page visit, and purchase

2. Opened broad desk&mob campaign with $0.2 bid and $250/day budget.

3. Our CPA is $55 and we get a sale from every 30 visitors who visit our offer page (we dont count the CR from lander/advert as it varies from every traffic source, so it'd be a leap-of-faith)

4. So we have 1 sale for every 30 offer page visits, and 1 sale for every 10 checkout page visitors (before retargeting, but we try to break even on front so can't rely on that).
Bottom line, we can pay:
~$1.8 per offer page visit
~$5.5 per checkout pagw visit

5. Every day we excluded all pubs who'd:

*spent more than x4 offer-pg-visit-cost (we set it to $8)
*spent more than x4 checkout-pg-visit-cost (we set it to $20)
*spent more than x2 CPA, $110 (lets say they gave us $2 offer-visits and $6 checkout visits but overall was spent $110 with no sale, we'd exclude the pub.

6. We kept A/B headlines and creatives. We'd run one image with 5 headlines, and then winner with 5 different images, then the winner with 5 different headlines, and so on...

Conclusion: we lost $5K, and barely got 10 sales (x10 on CPA). I can't accept the fact that it CAN'T work on native, and I'm absolutely sure that there's a way to crack this funnel.

Those are the main questions at the moment:

1. should we stop it and seperate the campaigns to mob & desk?

2. OB's targeting is any good? Maybe we should stop the broad campaign and try to play with targeting

3. Should we replace the strategy and go with $0.6-$0.8 cpc and lower budget to get higher quality traffic?

4. We never tried Taboola, but I guess OB pubs are quite good, as I see on all spytools similar campaigns that manage to scale
-----------------------
Guys, if you have any suggestions, tips, or critics, please let me know! It can really help us and I promise to do two things (call it Karma or any of those BS):

1. If I'll figure it out, I'll share it with STM community

2. Once the offer stops working (but assuming we'll crack OB and scale it), I'll share the entire thing (we screenshot, video, and record EVERYTHING)

Thanks for the help and have an amazing day!

P.S - To those who wonder, the CR on the advert is 7% as of today, while on FB for example, we get 35%. But that's ok, we also pay much on each click on FB.

---

Disclaimer: I havn't run much on ob so hopefully someone else can chime in as well.

-Definitely split desktop and mobile. If you look at your stats I bet almost all your traffic was mobile
-If you are running tier 1, $0.20 on ob is way too low. $0.60-$.80 is more appropriate

Sounds like you have a great offer/funnel. Good luck!

btw what do you think of Liveintent?

---

Sorry for the offtopic, but you share some details about Liveintent.

Do you maybe know if they also EU inventory?

Thanks!

---

Trick to native is first of all your creative, to dial in the low CPC in my experience.

---

Some quick pointers:

1. I'd focus on Taboola 100% and leave Outbrain for additional scale. More volume & better quality, campaigns scale faster + smartbid rocks. For a way smaller, but higher quality source to learn native on I recommend PowerInbox

2. Always start bids high, eg $0.75 for desktop then lower as campaign hits daily cap, until ads 'take' in the system.

3. Separate campaigns per device and bid level, eg like so:

Desktop campaign $0.75
Tablet & Desktop campaign $0.60
ALL devices campaign $0.50

4. Regarding blocks, I would be a bit cautious with the overzealous automated rules and recommend first running things manually. With many advertorials, clicks sometimes happen a minute or two after the fact and if you're launching a campaign and getting blasted with traffic you could be blocking some good publishers right off the bat.

5. Ads are the biggest lever on native so focus on that moreso than micromanaging publisher bids and blocks.

6. 7% LP CTR is bad, you should be seeing at very least 20% + for desktop traffic. It is an advertorial correct, written in the style of similar ecom offers you see live on native?

---

Fascinating post... especially about all those other networks... I agree- would love to hear more from you about your experiences with LiveIntent, Quora, etc!!

I think @thedudeabides perfectly summed up the important points to take note of.

I would add that on native, you have to remember, your ads are being shown on hundreds if not thousands of different sites, so its normal the quality of traffic on each of those sites will range dramatically.

The example I always like to use for this is Breitbart vs Conservative Insititute on Revcontent. These are two very similar sites, built similarly, similar widgets, similar volume, similar political orientation, similar demographics, yet for some reason ConservativeInsititue typically converts 5x better than Revcontent (at least for me and the other folks I've talked to).

It's likely your 5k went mostly to blocking all the bad sites on Outbrain and that you got very little traffic from the good sites yet, especially if you were combining mobile and desktop.

I agree with thedude that you would have far, far better luck running on Taboola SmartBid. In that case the algorithm will be doing the work and targeting people it knows will convert (as it's based on tons of variables down to the individual person clicking the ad, a la Facebook, rather than Outbrain or Revcontent where you're just bidding per site and nothing else typically).

Try it out on Taboola Smartbid and I bet you'll have great results! Just remember to try tons of different ads until you find the one that clicks and is both high ctr and high cvr and thus profitable for you.

And please do share your intl on those other traffic networks if you get a chance!

Good luck!

---

No idea, I run US&CA

---

Thanks

1. we'll copy everything to Taboola now and split DESK & MOB.
2. the rules are made manually for now, so we don't exactly pause it the second it goes out of goal cost
3. powerinbox was great for us but at least for us, it was a bit hard to scale (on $200/day it works like magic).
4. yes, it's the same advert as all those DFO/Giddyup offers but we'll obviously A/B it to get to 20%

---

Since you're already having some success with this offer on other networks, besides the above suggestions, I would suggest considering the following as well:

- Implement both Outbrain and Taboola pixels on your landing pages so that you can gather some audiences from your currently performing campaigns.
- Once you've accumulated enough visitors to create look alikes on Outbrain or Taboola (for Taboola you'd need the visitor's email addresses), you can start campaigns using this similar audiences.

In case you're using a click tracker for your landing pages, adding GA won't hurt. At least you'll be able to see the average time on page, bounce rate, etc of the visitors coming from a specific site/widget/section/ad. On top of that, you can as well utilize GA data to explore your campaigns demographics (as Google recognizes them).

---

Few questions about Taboola:

We run DESK / US -

1. Should we do any targeting at start?

2. Start with 0.75 bid and let's say 250/day budget?

3. Start with smartbid?

4. Objective ONLINE PURCHASES?

5.Her'es the rest of settings:

---

Settings look fine and yes online purchases. Open targeting at start. Smartbid always. 250/day is a good spot to start. May overspend up to 2x initially depending on how good your ad CTR is.

---

Thanks, appreciate the help!

I'll give it a month.

---

I can just upload CSV with 10K emails and run LAL on this file?

---

Hi man, very interesting thread. We have some ecom products doing very well on Native and we can help you run that profitable on TBL/OTB. Hit me up if you are interested! My skype is: nadja3193

---

If you have a list with 10K emails then Yes, you can use it to create LAL audiences for Taboola.

---

Thanks

---

Home > Paid Traffic Sources > Native