Seed Audience & ROAS (10)
Quick question:
Dose a seed audience have to have a positive ROAS for you to consider making a LLA from it?
For example, in gaming I am getting a ROAS of $6 - the game payout is $3 However if I am paying $6 on a very broad targeting, wouldn't a LLA on just those that converted for $6 yield a much cheaper cost per acquisition? In theory I would think because now I am targeting only those people that clicked on my ad + those that converted. So in theory wouldn't my link clicks be cheaper as well as my cost per acquisition?
Just wondering if I need a positive ROAS on a general target audience to even consider making a LLA... and if not how close... break even?
Simple answer: No
Long answer: No, because your cost is high down to the broad targeting. If you use an LLA (assuming you have had 300-400 pixel fires) - your audience will be less wide, and your CPA will drop significantly.
Right that was my thoughts exactly. Question is are there any averages? For example if I am paying $6 to $9 per lead right now and the payout is $3 is it worth collecting 300-400 of this and taking this massive loss in the data collection process in hopes of that with a LLA of 300 people who converted my CPA will drop below $3?
That really depends on a lot of factors.
I don't really know your internal metrics such as your funnel/ad ctr/relevance score/cpc/backend CTR - these factors hugely dictate your performance.
Your initial question was whether a lla needs a positive ROI for it to be worthwhile - the answer to that is no, but the lla will only help if your above metrics are somewhat decent already.
As a general measure (since you are doing gaming/cheap conversions):
- Ad ctr should be around 5-8% for more aggressive approach, 3-4% for less aggressive, depending on how BH/WH you are
- Backend Funnel - your offer should be converting at around 20-40% for such a cheap offer, possibly more
- Relevance score - this should be above 8, effectively if it is below 8 right now, facebook is sending you junk audience and you are being outbid by your competition.
These figures are a ballpark estimate I must add, and by no means concrete. I really have no idea what you're running so I've just given general benchmarks.
You need to measure the above, and optimise your campaign to really make an LLA work to your advantage. Simply put: the lla alone won't guarantee you profitability, you need to look at your funnel too.
Great feedback from kasher9. If your cpc is already on lowest point, your prelander CTR highest and you have the best converting offer, -50% ROI is not worth to proceed. But I am sure you can gain profit by optimizing your ads+targeting (less broad), increase CTR on prelander and perhaps find a better/similar offer and split test.
No matter how much you loose on creating a proper LLA, it's worth a lot, as much as a LLA which gained from being in green. BUT, the lower payout/cheap conversion, the less impact a LLA actual has from my experience. Unless it's really niche product ofcourse. A broad audience/product could potentially hit "anybody" within a certain sex, age, demographic range.
For example I never use a pixel or LLA for sweepstakes (lead gen, pin submit) since I "know" all women 35+ would love to win a certain supermarket voucher. Ofcourse relatively a small amount will never fall these kind of offers, but the majority does if you are convincing, trustworthy and clear regarding/during the whole funnel. I am sure I can gain more profit by using both pixel/lla eventually, but I prefer to put my time in optimizing the actual funnel (ad, prelander, offer page). Others prefer to dig deep in analytics and precizely cut and optimize placements and other variables.
Yeah right now I have made 11 videos for this specific game and tested 37 audiences, each adset it's own audience. No matter what I do my relevance score for this is always 4-5 on these videos. Even when I then tested 8 images after the videos just to see and the relevance score was low... which is odd b/c the targeting is laser. MMO game, targeting other mmo's in the same niche one by one. I think a better approach here might be to run this on native first and make a page that looks like the game and collect the emails and upload to facebook to start with a better seed audience.
That won't help your relevance score.
Your low relevance score is to do with front-end metrics; i.e. your ad ctr, your feedback, your comments, your likes, and overall performance against your competitors.
Trying to create a laser targeted audience via native will not help your relevance score, I can assure you of that, not enough to make it a 9-10 anyway.
What you need to try is different ads, and different ad formats, so perhaps try images too. And you may be bidding for the wrong placements; with games it's generally wise to bid mobile-->newsfeeds only (IG does well too).
A good way to split test is:
- 3 ads in one adset @10x your payout
- 4 adsets in one campaign
That will give you a total of 12 different ad variations tested, within 1 day. Keep doing that until a time you see decent relevance scores.
Just a hint: Your lander really makes a difference (facebook's AI is smart enough to match audience against your lander) and your ad copy makes a difference too. But I am 100% sure your low relevance score has nothing to do with your audience and everything to do with your competitors performance.
Originally Posted by kasher9
Originally Posted by stickupkid
Yes targeting can make a difference ofcourse, but in the OP's specific case he said he is targeting gaming audience, which is why I responded the audience will not make any difference to {his} relevance score.
However; why does the broad targeting strategy work for almost all BH campaigns? Not everyone is interested in selling their kidney to make money online (despite what affiliates believe). Facebook's samples the correct audience based on your ad copy and lander, that's why broad targeting works.
When you're playing around with 4-5m+ worth of audience, you will find that 70-90% of people are not interested in your crap, yet facebook still manages to squeeze conversions, and that is wholly down to your ad copy +lander, and nothing to do with your audience definitions.
This is also one of the reasons why many bh'ers fail - their safe page is not congruent with what they're trying to sell, and their copy is way too click-bait/generic ; ultimately this leads to poor quality/bad ROI.
Great conversation here. Just wanted to throw my two cents in on a few things.
I wouldn't necessarily worry too much about hitting CTR or CPC benchmarks. They're a great indicator of your ad performance at the surface level. If you're optimizing for traffic though, Facebook is going to show your ads to people who are most likely to click. So you can have a 20% CTR and dirt cheap CPCs, but your conversion rate won't be great - perhaps nearly non-existent. Another example is broad targeting. If you target super broadly, you may have a low CTR but your CPM and CPC could be low enough to offset your CTR. I'd stay focused on conversion rate and ROAS while slowly working to improve upon the rest.
Dose a seed audience have to have a positive ROAS for you to consider making a LLA from it?
Nope, the goal of your lookalike audience is to create an audience of people who are most similar to the people who have converted. Ad performance metrics wouldn't be taken into account when building this lookalike audience. You definitely want to include everybody who converted.
For example, in gaming I am getting a ROAS of $6 - the game payout is $3 However if I am paying $6 on a very broad targeting, wouldn't a LLA on just those that converted for $6 yield a much cheaper cost per acquisition?
Lookalike audiences are powerful, but 300-400 conversions may not be a large enough dataset to outperform other targeting options. Lookalike audiences aren't a sure thing until you're closer to 1,000 conversions. It's absolutely worth a test though. It's just one of those things that you can't know until you try.
which is odd b/c the targeting is laser. MMO game, targeting other mmo's in the same niche one by one
Laser targeting doesn't always work out. In my experience, it;s working less and less over time with Facebook ads. Have you tried combining all of the different games into one ad set and letting the algorithm do the rest?
yeah right now I have made 11 videos for this specific game and tested 37 audiences, each adset it's own audience. No matter what I do my relevance score for this is always 4-5 on these videos.
Did any of those audiences make it out of the learning phase? What size were they? How much budget did they have? (If you don't mind answering)
Home > Paid Traffic Sources > Facebook & Instagram