Running multiple accounts on the same demo (18)
So you're promoting offer xyz to demo xyz and you have multiple accounts in your arsenal. Do you keep 1 account per 1 demo or start outbidding yourself?
It doesn't work out to run multiple accounts on the same demo. What you will achieve:
-- higher click costs
-- lower CTR on both accounts (see above)
-- inflated reporting rates because people see your ad too much.
If you use different ads, it's the same lander and performance drops massively.
If you use a different lander, it's the same offer and performance drops massively.
If you use a different offer in the SAME niche, you will get kicked off one offer quick because the up-take rate drops AT LEAST HALF on both offers (assuming they convert the same). Nobody is going to deliver a good lifetime value to the advertiser if they're signed up to 2 offers.
If the offers are in a DIFFERENT niche, you are doing ok, but click costs are still inflated unless both your ads get clicked at the rate you can achieve if you just ran one at a time. The advantage of us Stackers launching ads is that the CTR is usually higher than regular businesses, but when you compete against yourself or another marketer in the same demo, it drops.
Maybe I'm not convincing but I can say I've tried a lot of things. Simply exhaust spending on one account. I've talked to the biggest guys about this. We've tried everything. Tried doing different targeting (splitting the geo states/countries between the accounts = drop in reach and increase in prices for both accounts), offsetting ages (24-55 in one account, 29 to 59 in another = the bulk is still being duplicated), even as advanced as when clicked on account A's ad, pixel activates and the user is added to the exclusion list for account B so FB no longer shows those ads. And a few other things that are a bit more shady. Nothing has worked.
One account per demo, but exhaust it and then move on. You'll get banned before you truly exhaust it. Good luck
Hmm, this totally depends on the demo if you ask me. Normally I agree with everything redrummr says, but this time I have to disagree ;-). If you target a broad audience (with no interest targeting) why would you NOT use multiple accounts? Your competitors are already competing with you, so why not take an extra ad spot yourself too? Sure, you're competing with yourself for a little, but it doesn't add up to the total volume on that demo.
Obviously have a different ANGLE on each account. But there's no reason to not promote the same offer, since others are doing it as well most likely. Back when dating was being cloaked a lot on FB, you'd see 3-4 ads all promoting the same offer and there is a reason for this.
Hey scito... this issue frustrated me to no end.
On POF, Bing, Adwords, competing against yourself is good. For some reason on FB, it's a bit different.
The algorithm sucks, combined with the fact you can't do day parting. On POF you can do tiered bidding, but this doesn't work well on FB at all. You just end up pushing your prices up.
FB algorithm seems to create links between advertisers and audience who have previously clicked on links. This can't even be gotten around by bidding for REACH (to reach new people).
In the past, we have already established that FB uses a disproportionate amount of clicks in the first 12 hours of the account. Therefore sometimes you need an account in another timezone to effectively hit your market at all hours without tweaking too much).
Let me throw a few thoughts out to illustrate how I've tested in my group:
In a demo of 5m, you can advertise in 1 account for 1 day, and see that 15% of your 4,000 visitors clicked twice on your ad. That is a 15% drop in performance (from what would be ideal, 1 click then they don't get to see your ads again, as they are most likely to convert the first time (what my stats tell me), we all live with it. FB doesn't do click capping (I believe!), and I have tried to implement this with a method of "click --> lander has hidden iframe/JS safe method that loads audience pixel --> audience pixel is an exclusion zone in this current campaign automatically so user is not shown ad again from this account". It doesn't work, because the audience pixel is effective only when first launching the ad. It's good to use this strategy every month (use your audience pixel) so that everyone who clicked LAST month does not see your ad this month. THEN you will get all new clicks for this month (respective to last month, but still get duplicates from this month). But unless you want to make new ads every single day, you will actually get the same users, where possible, over and over. Facebook likes it this way, it makes ad serving easier.
Add one more account on top, and 15% becomes 60%. I have tested specifically with one account in Perth and Sydney timezone (3 hours apart), not linked, 100% different, and when advertising to ROUGHLY the same demo (changed ages a bit, changed bids slightly to enter a different 'stream' of bidding according to the FB documentation), the unique clicks were maybe 40% of the people who clicked. I've repeated this. So in total, out of let's say 4,000 clicks, a full 2,400-ish of them were not one-click-only people. The people who clicked the second time were NOT any more likely to convert -- they simply cost twice as much CPC. I advertised with different accounts and ad angles (pictures, overlay text in the ad + FB Page was different due to being a new account) and FB delivered the different ads to a set of a few thousand receptive ("ad-clicky") users in both accounts.
I then switched bidding styles, and kept getting similar drops.
There is more to talk about, such as an advertiser seeing the same IP delivered by the same affiliate in the same day = they will scrub hard proactively (think you're sending rubbish traffic), have had this happen on two networks and had to ask for new affid to be made on one network. To get around this, you can have multiple signups to the same network (not really going to happen) to emulate being different affiliates (though this is the best way to get around it and emulates your scenario well), or use different networks (it's likely performance suffers on other networks, which is why you chose the network you chose).
Angles don't matter. If the angle is different, people click more, then you have even MORE duplicate clicks, and more reporting. If the angle is the same, people don't click, your prices skyrocket. LP CTR only goes down from 24% (flog) and 54% (dating) to 8% (flog) and 28% (dating) respectively, when trying to do things on 2 accounts instead of 1. I kept a trail of what I did trying to solve the issue.
If you are advertising to a large demo, check your stats. When you look at it over a WEEK, you will see that, from all visitor IPs, 70% are NOT one-click-only visits. Over a month, it will be 75% or even more. FB has unique users, but there is no click-capping and for the life of me I can't emulate it, though I've tried everything. I've tried approaching it fresh every time.
The best thing IMO is to use a single account and spend high to reach more users, rather than using multiple accounts. If you have done the checks and not seen a performance drop I am astounded, scito, and I gladly bow before you for any crumbs of knowledge you can share :-)
Long and thorough explanation and I fully get you. I dived into my stats and I'm going to share them here. I first want to address a few points from your post which should not be seen as an obstacle from my point of view.
- Regarding scrubbing from an advertiser: this is why you need to know who you do business with. I currently only have direct deals running and I'm 100% sure there is no shaving going on since I have access to the direct CRM systems.
- Multiple signups per network: this is possible, depends on the relationship with the network of course. I'm not going to put down names here, but there are 2 networks where I have 3+ accounts each. Obviously networks aren't going to openly talk about this.
That being said, I took stats from 1 campaign which is spread out over 4 accounts (2x PST, 1x EST & 1x CET timezone). Timeframe: september 1 - 26. Every account has a different angle and every account has a slightly adjusted LP geared towards the angle of the account. The total demo I'm targeting is 2.2 million. That being said, I ran the following queries in phpMyAdmin on the mt_click table since I use iMobiTrax. Here are the results:
Account 1:
Total clicks: 41,678
Total unique clicks: 31,204
(25% duplicates)
Account 2:
Total clicks: 22,094
Total unique clicks: 17,773
(20% duplicates)
Account 3:
Total clicks: 22,251
Total unique clicks: 17,864
(23% duplicates)
Account 4:
Total clicks: 16,809
Total unique clicks: 13,577
(19% duplicates)
Total amount of clicks over all accounts: 102,832
Total unique clicks over all accounts: 71,541
(Total unique clicks was taken with following query: SELECT click_ip FROM mt_click WHERE camp_id = '1' OR camp_id = '2' OR camp_id = '3' OR camp_id = '4' OR camp_id = '5' OR camp_id = '6' OR ... GROUP BY click_ip)
So total duplicate clicks over 4 accounts is 30%. This is all from desktop traffic, no mobile traffic. Your statement of "you will see that, from all visitor IPs, 70% are NOT one-click-only visits" is not correct according to the statistics I have here unless my data is not statistical significiant.
I actually did the same statistics for a mobile campaign on a 1.4mil demo spread over 3 accounts which I started 8 days ago.
Total amount of clicks over all accounts: 33,487
Total unique clicks over all accounts: 25,937
(23% duplicates)
Perhaps it's too early to tell, but it seems like click loss is lower on mobile, but this can also be due to the fact IP's change a lot more on mobile devices so 1 user can still have clicked multiple times without you knowing because the IP differs.
I'm a strong believer that FB assigns your account to a specific "group" / "audience" once your first data is rolling in from your first campaigns. There is a reason they want you to use 1 account per CLIENT and I believe this is why. By taking a unique angle / ad style on every new account, you get assigned a new "group" / "audience" of potential click happy people. This is just my theory though.
It's great to have a dicussion going about this though, I feel like we can have a few eye openers if we work together. What do you think about my stats?
scito,
Are you running white offers?
Thanks for contributing, scito. Your data is interesting to me; it does not match with mine, but nonetheless thank you for posting it! There needs to be more in-depth discussion.
I feel that the amount of truly deep discussion is limited due to people's FB accounts getting slammed, and nobody worries about the finer details of advertising on the platform. Because they would be more worried about the meat of their success: keeping the account active. For me, I know FB is a beast and it just takes ONE account and ONE month to make me a happy man for a while.
Gents
thank you so much for this high level discussion of the merits / down sides of multiple accounts as the spending ceiling is a real issue for me.
Its nice to see the level of thought that goes into it and backed up by DATA not opinion
thanks
On a side-note, with the deprecation of grey accounts and the ability to make new ad accounts personally through the business manager, perhaps things will become slightly easier?
I'm not sure if the new accounts will be vetted or sand-boxed but it is nice to have the ability to make client accounts at will rather than having to go through a rep.
This might also make it easier to test things like this.
By the way, when I talked to my rep at Social.com, he said they ran a test with a very large advertiser - over billions of impressions, and found that the single advertiser did not adversely raise news feed CPM prices. In other words, they were spending huge numbers on broad NF impressions and didn't seem to increase the pricing.
When you have multiple accounts, you will still be limited by the frequency capping at the brand/page/ad level.
So, in your tests, were you using different pages? That would circumvent frequency capping and you would then likely compete with yourself. Given that the aforementioned large test showed no apparent price increases, this could be due to the CTR changes rather than CPM increases (any data there from oCPM bidding?).
Quite a complex ecosystem, no?
Not tested yet, but new ad accounts would probably still require the same verification and be subject to the same security triggers... It's yet to be seen whether getting one of these accounts banned also gets the parent Business Manager account banned.
Originally Posted by redrummr
Originally Posted by bogeyguy
Originally Posted by redrummr
Originally Posted by zeno
Originally Posted by scitox
Originally Posted by radiosurf
I think Scito means that FB watches new accounts and e.g. sees they run gaming ads > classes account as gaming, run weight-loss ads > account flagged as weight-loss = monitor relatively closely >> influences likelihood of retro's etc. later on.
I thought he mean't that FB sent traffic to only a "specific "group" / "audience" out of the total demographic, but I must have misunderstood.
I'd still like to understand why Frequency starts becoming >1 when only a fraction of the total demographic is reached. For example, if the demo is 500,000, and an ad has only been shown to 25,000 (Reach), why does FB start showing to the same 25,000 causing the Frequency to climb, rather than show it to the other 475,000? Possibly because only 25,000 are active?
Not everyone logs in every day.
Your audience is not homogeneous in the ads they are delivered - they will belong to different audiences and be classed differently in the eyes of the CPC/oCPM/CPA algorithms, so there impressions will be worth a varying amount = some users cheaper to reach than others.
The bidding mode used influences the subset of your total audience that you will reach (at least initially).
Your bid/budget will effect your reach. Low persistent reach, IMO suggests bidding too low.
Thanks, Zeno.
That ad is oCPM at the default. The budget for that ad is $100 a day, compared to only $20 per day for the other oCPM ads that aren't getting throttled. That's why it's seeming like it's related to the domain.
It could be that the audience subset for actions is quite small for your target audience.
Try a duplicate ad that bids say $1 for clicks (rather than $5 for actions, the default I presume you were on?) and see what happens.
Home > Facebook & Instagram > Account Issues and Questions