Home > Paid Traffic Sources > Facebook & Instagram

---

How much can you spend per day on FB with an audience of ~2 million (16)

---

Hey guys,

let's assume I have a profitable campaign where the only targetting is age and gender. The total size of the audience is arround 2 million (data from ad manager).

What is (roughly) the maximum possible spend per day? Maybe arround 2000€ ?


From my understanding I either need to increase my bid or increase my CTR to get more traffic. I also know that the top players on FB mainly use CTR to get more traffic, not a higher bid.
But assumming that I can't increase my CTR any further ... I'd need to increase my bid, and that would mean that a 100% ROI campaign would be only breakeven when you have to double the bid, correct?

Ads are already shown with a frequency of 1.6 and only 500k reach last month. The CTR of the top ad also dropped from an average of 2% to 1% (newsfeed), allthough banner blindness can't be a problem yet (not enough reach/frequency) and the banners have not been ripped. Funny enough, while the CTR went down the CR went up and the cost per conversion dropped only slightly. (I bid CPC, so there should be no optimization from facebooks side at work).

And is there a rough estimate how much the effectiveness of a campaign (not an ad) will go down with higher frequency?
How do I scale? Just slowly increasing daily budget by 100$ per day, keeping the max bid 1 cent over the average cpc, and replacing ads when their CTR drops due to higher frequency?

Okay, my main question is:
- max spend per day with 2m audience? (roughly)
- how much more expensive will a conversion get if I want a lot of traffic? (rougly)


Thanks guys, appreciate any advice.

Cheers
Steve

---

Perfect world scenario:

1/2 of that 2 million demo log in every day = 1,000,000 potential reach = 1,000 mille
Can therefore spend $1,000 x effective CPM

If your demo's average CPM is $5, that's a maximum of 5,000.

However, it is virtually impossible to reach every single user with your ad - the news feed simply doesn't have enough ad inventory for that, so it's not going to happen unless you bid veryhigh.

So, you might only reach half of these people, giving $2,500 daily.

At the moment you are bidding with a CPC cap... this somewhat limits you - if you bid 1c over your CPC you are creating a ceiling, and if CTR is invariant, this creates an eCPM ceiling as well. As a result, FB is not going to serve your ads over the top of others bidding oCPM who can generate a higher eCPM for Facebook.

If you want higher position in the news feed and a greater penetration into your audience, you need to bid high and let Facebook charge what they want.

There will be an optimal position for you where news feed position, ad costs and ad revenues are maximised.

Frankly, this is where oCPM comes in - the goal of oCPM is to optimise toward this maximum using CVR data.

Right now though, I would raise the ceiling on your CPC bids and see what happens. For all you know, it might only take a 15% increase in actual CPC to raise your eCPM enough - but sadly I doubt this is the case given the relatively low (1-2%) CTR of the ads.

As for your specific questions...
1. Estimates like this have been made before but mainly for RHS ads - i'm not sure frequency in the news feed has become a big problem yet due to the limited ad inventory and generally low per-user impression counts.
2. That scaling system sounds fine, though the CPC ceiling might limit your market penetration or daily reach purely due to the eCPM ceiling it also sets (which drops further as CTR falls).
3. An example calc I would do: (Log in %) x average CPM / 1000 = spend. E.g. 50% x $5 / 1000 = $2500. Assumes perfect conditions where every user sees your ad i.e. your eCPM is the highest of all advertisers.
4. There's no correct answer here... but if bidding CPC, assume the demographic is homogenous i.e. as reach expands, audience CVR remains steady. The increase in bids/CPCs required will depend on the competitive landscape, which isn't something anyone can predict - you might need to increase bids by 10%, maybe 100%, and this could vary with the day, time of day, holidays, ends of fiscal quarters, etc.

---

A few observations/questions related to this interesting discussion:

--What percent of potential audience do you really want to reach each day? Is your offer really strong enough to blast 50% of the potential audience every day?

--When FB says "reach", do we know what time period that is over? I notice frequency goes over 1 almost immediately, seemingly indicating FB can start showing the same ad to the same people quickly, which could lead to ad fatigue and lower ctr.

---

Interesting thread, thanks for that helpful reply Zeno.

I'm still learning how to scale on FB and posted about it before, redrummr had some great input: http://stmforum.com/forum/showthread...l=1#post153263

---

When you optimise for reach presumably it has no temporal component, i.e. it just tries to reach people as much as possible for each opportunity, even if it has to be very very low in the news feed and multiple times per user per day.

---

First, thank you for your detailled answer.

Wouldn't that be true if and only if the campaign frequency is exactly 1.0 per day?
So half ot the userbase would see one of the ads of the campaign exactly once per day.
But it would be very easy to say that the frequency could be significantly higher like 3.0 per day or a lot lower like 5.0 per month.


oCPM sounds like a good idea in that context, I'm just so used to CPC and most people who run higher budgets on facebook seem to use CPC.
But just entering the net profit of each conversion goal into the ocpm seems like a very good idea - let facebook do the work

So the short answer about the max daily spend was: arround 2500$, definitely more than 1000$, definitely less than 10.000$ / day

---

I heard somewhere that 1k per million was a good guideline and have used it with decent results, so in this case 2k a day.

---

Yup, clearly quite a few factors involved, and assumptions that frequency won't change CPC costs or CPM, and many others no doubt.

If bidding oCPM remember it is a CPM bid so don't go putting in $30 haha. Run the defaults or start with e.g. $5 and test from there.

Also note if using multiple bids i.e. $2 for reach, $5 for actions, the average CPM FB will bid is the sum of all of these. It would be an expensive lesson to put in 5's across the board.

Yesterday I upped some oCPM bids to get over a traffic bottleneck and went from ~$30 spend / $24 revenue to like $130 spend / $45 revenue. One of those typical FB-induced GOD DAMMIT moments.

---

So to summarize, scaling should be relatively easy on facebook and going from 400€ / day to 2000€ / day spent should be the easiest part of a campaign, right?

And everybody recommends oCPM for spend-optimization? Damn, I got so used to CPC by now

---

oCPM is more volatile but almost guarantees volume.

I wouldn't discount CPC at all, I just don't start with it anymore, and it removes the ability to let FB target audiences with higher affinities for certain actions.

If you have a campaign at $400/day spend then yeah I'd say the jump to $2k is less of an issue than the jump from $40 to $400.

---

There is no capping on the offer (we're the offer owner), if that's what you meant with "strong enough". The maximum would be determined cashflow wise (there are 2 weeks delay between ad-spend and revenue).

Not sure if I get that correctly. The time period is just what you select - if you choose the last week, it shows the frequncy of the last week. What is does not show though, is how the average frequency is complemented, like which percentage saw the ad 1x, 2x, 3x, 4x etc.


Just to clarify: "temporal component" means in this context that the ad doesn't "remember" anything from its past, correct?


Interesting!


Clarification: You wouldn't enter the real revenue per sale in the oCPM specification but something lower? Wouldn't that discount the use of oCPM to maximize profits (not ROI or revenue, but profits) ?


Thanks for all replies.
Sorry, if I'm sounding like an ultra noob here.

I can't really use oCPM if my conversion goal comes 10 days after the ad click, right? I'm using an autoresponder sequence. The delay between the ad click and the sale will probably kill the optimization.


Are the following assumptions correct:
-- generally speaking, ROI goes down when you scale up (as you need to bid higher to get more reach - allthough you might also get more priority in the adserver due to the higher budget, not sure if the last one is a factor, it is on some platforms)
-- ROI doesn't go down extremely on large audiences, so a 300% ROI campaign on 300€ per day is likely to be scaleable to 2000€ / day with still over 200% ROI on average.
-- do use multiple ads and different angles at once to avoid offer saturation
-- find a balance between ads with good ROI and userfriendlyness to avoid ad reports and possible disapprovals/bans - same with frequency, if users see an ad too often, they'll probably be a lot more likely to report it

---

Just to clarify: "temporal component" means in this context that the ad doesn't "remember" anything from its past, correct?

By that I meant it just tries to reach everyone, every day, regardless of what happened yesterday. In other words, it doesn't plan to 'reach everyone with 7 days' or anything like that.

Clarification: You wouldn't enter the real revenue per sale in the oCPM specification but something lower? Wouldn't that discount the use of oCPM to maximize profits (not ROI or revenue, but profits) ?

Actually, I'll have to clarify this for you sometime later in the week.

Initially, Facebook said that you bid something comparable to what that action is worth to you - i.e. if actions/conversions give you $10, you bid $10 and let the system try to optimise to bring positive ROI.

Facebook initially allowed relative oCPM bidding but have moved to absolute CPM bids. In a recent test, I found I got no traffic at $5 action bid, and upping it to $10 gave a burst of traffic, I then brought it down to $7. The CPM bid that I am actually paying is averaging to around $7, which suggests that the actions bid is determining average CPM bid - but, this could be chance.

On the developers page, it says the oCPM values are not 'bids'. However, via a PMD, I am told the oCPM bid(s) sum to the average CPM for the campaign.

This is very confusing I'm sure. I have a call with a PMD tomorrow so will clarify. For now, assume that putting in the value of a conversion is fine, but I strongly suspect these are now akin to CPM bids, i.e. what you are willing to pay for 1,000 impressions, not for a conversion.

As for your goal optimisation, use a conversion pixel for when people sign up for your mailing list - if you have control over this then for god sakes don't use a pixel that fires 10 days from now!!

1. Don't make assumptions.
2. Don't make assumptions.
3. Always test multiple ads or don't bother marketing at all. Once you filter down to the best, if they are within similar profitability of each other, I would usually leave them all running for the sake of staving off banner blindness.
4. Affiliate marketers always live on the edge, just don't risk your account recklessly. You can't really control frequency, but can be weary of how much you can spend daily relative to your audience size. If frequency is high you probably want to reduce budget not for fear of reporting, but because ad performance will likely degrade. If people are likely to report your ads, you have bigger problems than frequency.

---

Thanks again.

A different calculation approach:

2 million audience
Assumption: You can convert 1% of that audience into leads (I just made up the 1%)
That's 20.000 leads
Let's say we have a profit per lead of 4€ after adspend.

-> Then that's just 80.000€ profit. That can't be right.

Are there any estimates how much of a market can be converted? (it's a mainstream dating offer).

---

Without some historical data, you simply can't make a prediction like this. And it would be precarious to do so.

1% of the audience? It could be 20%, but it might take a year to do it. It could be 0.1% and burn out in 7 days.

It really doesn't matter. Don't make assumptions about the end-game profitability/scale of a campaign, just roll out a test with the intent to scale as much as possible if it works. You can then make extrapolations and predictions from the data that is specific to that campaign and all it's inherent properties.

---

$250

$500 – most common for starters

$1k

$2k

$3k

$5k – The most typical bump spend limits when requesting from reps.

$10k – Risk management level

$20k – Risk management level

---

You posted account limits, this thread is about audience limits

---

Home > Paid Traffic Sources > Facebook & Instagram