Home > Paid Traffic Sources > Facebook & Instagram

---

First Facebook campaign, lead gen Campaign Tier 1 (24)

---

Hi All,

This is my first real attempt at a Facebook/Social media campaign.
Have been doing search for several years but have never really tried Facebook before.

So, Ive been promoting a Finance lead gen offer into a Tier 1 country for a while via search with some fairly good results.
The offer pays - 5 - 10 Euro per lead.
This activity has allowed me to build a Website visitors list in Facebook of about 5000 visitors.

From this website visitors data, I have generated a lookalike audience to 1% of the population
(choosing only 1% for now as I understand that this will keep the new audience as accurate as possible to my original website visitors list).

So, using this lookalike audience, I created a campaign targeting one gender using Mobile devices from the ages of 18 upwards.
I separated the ages out into adsets of every 10 years.
And each of these adsets then has 4 creatives which are simply the merchants banners
and my own text (slightly different banner and text for each creative) and these are simply direct linked through to the offer.

The traffic bidding strategy I am using is Link clicks to website, with a manual bid amount
set to £.01 below Facebook's suggested bid with delivery type set to Accelerated.

I set this campaign running on Friday 3rd and have run it until this morning (Monday 6th) and am now working through the results.

So, the overview is:



I don’t think that this is looking too bad to begin with.

And then going own into the age brackets, I can see some green for one particular age bracket (2).



Looking at operating system, it appears that IOS is the best converting OS but has lower traffic levels than Android.



And then looking at the Ad creative it looks like number 3 could be a winner.



My own questions now are:

1) Have I generated enough data so far for any of these results to be significant?

2) Would it be a good idea to try using a landing page so that I can add more
personalization in to improve the user experience? (currently I am just direct linking from banners straight to the offer).

3) How should I begin trying optimization and scaling of this?
My initial thoughts are to take the best performing combination of Age bracket,
Operating System and Ad, and try that combination to a bigger lookalike
Audience list (to say 5% of the population).
But as I said, this is my first Facebook campaign so I’m not sure if this strategy is a good one.

4) Are there any other metrics I should be looking at to optimise this one well (e.g dayparting/weekparting etc)

5) Would a different traffic bidding strategy be better to try for this?

Any thoughts or advice on this campaign would be appreciated.

Mark

---

Could you provide us some info on your ad CTR and CPC cost on FB? What relevancy you hitting? etc etc
Great first results by the way!

---

Sure, here is a screenshot of the biggest traffic spending ads.

My overall average CPC is £0.18 per click (0.02 Euro to keep it simple as I'm earning in Euro).
Average relevancy looks to be around 7 and unique Ctr is 3.23% with Link Ctr being 2.62% (not sure what the difference is here).

---

Unique per person - CTR = based on unique ip adress
Per impression - CTR = could be a person is seeing the ad more times than 1, before clicking on it

1) Have I generated enough data so far for any of these results to be significant?
Yes, based on provided results here you got some data to take learnings from

2) Would it be a good idea to try using a landing page so that I can add more
personalization in to improve the user experience? (currently I am just direct linking from banners straight to the offer).
Adding a lander which warms people up, make them excited, would definitely raise conversion rates. Assuming it's a good lander ofcourse!

3) How should I begin trying optimization and scaling of this?
My initial thoughts are to take the best performing combination of Age bracket,
Operating System and Ad, and try that combination to a bigger lookalike
Audience list (to say 5% of the population).
But as I said, this is my first Facebook campaign so I’m not sure if this strategy is a good one.
See if you can find out which images perform best, test it on low budget with several texts (maybe even a bit more agressive). I think you even need better images to reach relevancy 10, which means 2/3 less CPC cost on FB probably. I think narrowing down to the best devices, find better images and adding a good lander before the sales page would make your campaign overall in green! I wouldn't touch the target audience yet, since that 1% is quite accurate. If you widen it now and your new tests turn out to be shit, you not sure if it's your changes to the ad, or the audience. You might create a new LLA not based on web visits but on leads only, to be really specific!

4) Are there any other metrics I should be looking at to optimise this one well (e.g dayparting/weekparting etc)
That's why I asked for the FB stats, your cpc isn't bad, just like the relevancy. But imagine you can reach a 10 with a good ad. From 0.20 to 0.15 doesn't seem a big deal at first sight, but it's 25% more margin for you in fact, aiming for 0,10ct would be best and I think reachable....

5) Would a different traffic bidding strategy be better to try for this?
I prefer to do automatic bidding, just so I know for sure I get the "best" spots in quite a short time. Upfront I pay "much" on failing ads, but at least it gives me quick insights. I wouldn't change your bidding strategy now, just make sure you launch new ads in the morning/early afternoon, so FB has the time to optimize your budget without a rush during the day.

---

That is all very useful feedback, thank you

Another quick question. When I run search campaigns I find that I can hit some evergreen campaigns
that keep on going, sometimes for years as the demand just keeps cropping up (search = demand).
I assume that's not going to be the case with Facebook campaigns as I am showing my ads to
pretty much the same audience demographic and I guess people will just tire of seeing my ads
and stop clicking on them.

Is there a realistic time-frame that a good Facebook campaign can keep on running i.e weeks or months?

---

Depends on your daily budget, volume of audience, and refreshing time of ads/campaigns.
Try to find the right balance of maximum ROI and duration of campaign. It's very hard sometimes, people are greedy ya know!
But on a low xx budget a campaign might work for months yeah!

If you gonna push the same campaign to the same audience, you should think of different angles within that target group to attract in a certain way, eventually, all of them ofcourse. And keep refreshing images! Looking fwd to see your new results.

---

Okay so I got a little sidetracked with this one hence lack of updates, but on Friday i started
running it again over the weekend as I did before (so from Friday afternoon until today (monday) midday).

My main point of focus here was to test some new image ads to try and get the relevance score up
and also some landing pages as well as just going direct linked.

Here are the results.

So the overview:



On the face of it this is disappointing as its significantly down on the first time that I tried this campaign.
However, when I look into other results I can see why.

Looking at the ad performance.



Ads 3 and 4 here are the new image ads that I was testing. As you can see they spent a lot
of money and both had very conversion rate.
The relevancy score in Facebook did go up to a 9 for both of these ads but they CPC didn't really
improve. (still £0.18 per click).

I am thinking that my new test ads were a little to vague/generic hence attracting a lot of clicks
and improved relevancy but a lower conversion rate. This of course degraded my overall results.

And here are the results for the landers:



So overall for this time period direct linking is still performing best.
However, I realise now that I set Voluum to show the direct linked one
50% of the time and the other 3 landers had to share the other 50% so
not as significant data. Lander number 2 had a reasonable Conversion rate
and in fact when I look at the stats for today only (now that I have adjusted
Voluum to show all landers and direct link more equally) Lander 2 has the
best conversion rate at 4.48%

So my own little takeaways from this test are:

1. Images in Facebook need to be quite specific, just attracting clicks and
improving relevancy is not necessarily going to improve overall performance.

2. It seems that landers really can improve performance (looking at todays results, lander 2 is doing best)
but of course you need enough data to be significant. Maybe I am testing too many landers here (3 as
well as direct linked)? Therefore not getting enough data quickly enough?

So my questions are:

1) Any thoughts as to where to go from here? I am thinking to try some new images again, but
really make them specific as well as catchy to stop unqualified clicks that degrade the campaigns.

2) Regarding my lander tests, I want to do more here as I know (and can see) that landers can
improve conversion rate.
Would it be better if I remove the direct link test for now and just test 2 or 3 landers against
each other. Then when I have a best performing one go back and test that one against the
direct link? Is there a method/process for split testing landers?

Thanks in advance

---

Can you find a reason why ad 2 is doing good, and the rest not? Try to find similar images as ad 2, but also try totally other ones still aside from 1,2,3 and 4.
I would drop the landers for now and focus on the direct linking to be honest, especially if making landers isn't your best skill (no offense).

Try to find out why 3 and 4 don't convert that well, text wise. It seems you are promising some stuff in the ad what the campaign isn't offering. I know it's vague, but without seeing the ads/campaign it's hard to tell from here.

You target very specific right? Be sure you take out the risk of CTR/CR being low because of a too wide targeting.

---

Thanks for the feedback as always.

>>Can you find a reason why ad 2 is doing good?
Hmm, well it is very specific, both the image and text completely describe the offer.

I tried to be a bit more creative with ad images 3 and 4 (the text was almost the same as 2 though)
using a much less specific image that I hoped would catch attention (which it did but the attention seemed
to be clickers rather than buyers).

So yes, I am trying a slightly different version of Ad 2 to see if I can improve its performance.
And also something very different with a different angle also.

I will leave the 2 best landers in place for now and test these against the deep link. I am not
so used to creating landers, but I think its worth to continue the test until more data is obtained.

>>You target very specific right?

Well as said before, I am targeting a 1% lookalike audience based on 5000 website visitors.
Hoping to get this all working well and scale up to 2%, 5% and even 10%.

Do you have any thoughts though as to how I can target this even more specific?

---

Try to get buyers data, or at least lead level!

---

Ahh yes of course, that would be very specific.

Okay, so what is the usual method for obtaining such data? I assume I just contact the offer provider?

In the instance that they cannot give me such data I am collecting emails via an email opt-in on my
landers (so this would qualify as a lead I guess) and I then plan to make a lookalike list from this
data which I assume would be more targeted than the 1% lookalike of website visitors that
I am currently using.
At what point though would you advise that I have enough emails (leads) to create a lookalike
audience from it? I know that its the more the better but I could I try when I have say 500 emails
or should I have much more than this?

---

500 or more will do yes for sure! Ask yourself if the leads you are catching are likely to convert yes or no? If you attract them with "sign-up now for free xxx" you understand it will be not that good quality...

---

I am literally asking them : "For further information, please register your email" so I am fairly sure that they will be genuine leads.
Thanks as always for your feedback, I will report back soon.

---

Just another quick question regarding this campaign.

Up until now I have been using 'Link clicks to website' as the campaign objective.
Would it make sense to change this to 'Website Conversions' objective?

---

If you can place the FB pixel on the advertiser's page, yes definitely. How is your cpc/ctr/relevancy doin at the moment?

---

Hi again,

Okay so before I read your post about the FB pixel in the advertisers page I yesterday afternoon went ahead and tried a campaign
using all the same details (campaign, landers, ad creatives etc) except with Website Conversions set as the
objective (I paused my original Clicks to Website campaign while trying this).

Of course, I realise now that without the FB pixel in place, a clicks to website campaign cannot 'learn'
what is converting and so therefore cannot actually do its job of optimising for converions?
I do have the FB pixel in place in Voluum and it is passing back some lead data into Facebook, but it doesn't seem
to be able to attribute the leads to specific campaigns (I have another active post on this problem here:

https://stmforum.com/forum/showthrea...ok-with-voluum

So, the results from this test are:



-91% ROI Ouch (money values here are Euro not dollars also)

I am a bit confused though because as I said all was the same as my clicks to website campaign except for the
objective and letting Facebook autobid for me. The CPC for this campaign averaged £0.26 (0.30 euro) which is
a bit more than the clicks to website campaign but not crazily so. So its strange that the ROI results were so much
worse than my original campaign.

The relevancy was better on this campaign, some ads hitting 9. I assume Facebook is working to target
people more likely to interact with my ads.



But my takeaway from this then is that you should only run a Website Conversion objective campaign
if you have the pixel correctly attribting leads back to the campaign, otherwise Facebook is working blind,
is that correct?

I have paused this Website Conversion campaign until I can get the pixel working correctly.

With regards to my initial Clicks to Website campaign. Ive restarted it again and its ticking
along pretty much as before with a CPC of about £0.18 and ad relevancy of 8. By testing
some new landers the conversion rate has improved and I'm starting to see a bit of green
most days. Will work more on the ad banners to improve relevancy and will post a full
update soon.

For now, if you could give me your thoughts on what went wrong with the Website Conversions
test (I assume pixel not correctly attributing leads back to the campaign?)

Thanks,

Mark

---

Your CVR (and CR) now looks way lower than your first test. Did the advertiser change anything to the campaign or you changed campaign?
Good to see you found a way/angle to make your ads more relevant, this will result (with some patience) in a much lower cpc at FB.

Also it's way more difficult to keep the cpc low on FB than to find a good converting offer somewhere.
With the right pixel settings, some better LLA maybe and your new ads I think you are on the right way to make some profits!

Good luck with the pixel issue, unfortunately I can't help you with that!

---

Hi:

>>Your CVR (and CR) now looks way lower than your first test. Did the advertiser change anything to the campaign or you changed campaign?

Yes, as I wrote above, the results I posted here are for a 'Website Conversions' campaign. Previously I was running a clicks to website campaign.
The clicks to website campaign has pretty much the same results as before, so will continue working on and cultivating a more accurate LLA as
you advise.

But, my question again re my poor results above was:

"I should only run a Website Conversion objective campaign
if I have the pixel correctly attributing leads back to the campaign, otherwise Facebook is working blind,
is that correct? "

---

Apologies, clear. It doesn't make sense to run a Website Conversion campaign without the pixel played properly indeed.

---

Okay thanks for the feedback, I will work on getting the pixel allocating leads correctly as I think it can really help.

---

With regards to getting a more targetted LLA in place, I am currently collecting leads data (people who register their email in my lander).
When I have a few hundred (say 300?) I will try to create a LLA from this.

In the meantime though, what would be your thoughts of creating an LLA from people who engaged with my Fb page?
I have 1500 likes so far, is this worth a try? I don't hear much about LLA's created from page engagement.

---

Any thoughts on this please?
Would it be sensible to try creating a LLA of people who engaged with my Fb page (1500 likes). Or would
this just be targeting tire kickers rather than buyers?

---

On leads or sales are best at the end. Views, clicks and likes are too vague in my honest opinion. Likes could be interesting if it's really a "small" niche.

---

Okay yeah it makes sense, I will stick to gathering leads and then LLA from that.

---

Home > Paid Traffic Sources > Facebook & Instagram