Facebook Advertising Strategies for Early-Stage Startups(blog.interstateanalytics.com)
blog.interstateanalytics.com
Facebook Advertising Strategies for Early-Stage Startups
http://blog.interstateanalytics.com/2016/02/29/paid-user-acquisition-for-early-stage-startups-part-3-facebook-advertising-basics/
35 comments
Would you mind sharing some technical detail on how you used attribution as a backbone? What sort of modeling did you do around it? Just static attribution models? Or did you get into the dynamic algorithmic side of things like the big dedicated attribution platforms are doing?
Attribution on mobile is a commodity now. 5 years ago before AndroidID and Apple IDFA, all you could use was IP, OS, Device Type, and random other bits of info the network would send you. So you'd need some machine learning inference to connect installed users to where they came from. Now, you just literally match the device-ids with what you get in the click from network partners (off-Facebook). Adjust, Appsflyer both do this.
On FB, just use FBSDK and it comes with its own attribution. They are starting to shift to a "multi-touch" model where you don't really know how they attribute -- its a black box that FB controls via their statistics. FB will report what the CPI is to you. It sucks, but thats market dominance for you.
What I'm also saying is, make sure you're also sending revenue, logins, and any other significant events to FB as well. Then you can figure out ROI and do more performance marketing. This works at any scale.
On FB, just use FBSDK and it comes with its own attribution. They are starting to shift to a "multi-touch" model where you don't really know how they attribute -- its a black box that FB controls via their statistics. FB will report what the CPI is to you. It sucks, but thats market dominance for you.
What I'm also saying is, make sure you're also sending revenue, logins, and any other significant events to FB as well. Then you can figure out ROI and do more performance marketing. This works at any scale.
Thanks. I should have clarified--attributing an app install on Android to the source is definitely a solved problem. I was referring more to cross-channel attribution and determining dynamic weighting across multiple touch points.
Sounds like FB is working towards obfuscating that and not giving granular control which is disappointing to hear, but have you done any work around custom weighting like that?
Sounds like FB is working towards obfuscating that and not giving granular control which is disappointing to hear, but have you done any work around custom weighting like that?
The bit about attribution is critical, but misses a more important basic piece which is "make sure you setup and properly QA your conversion tracking."
So many companies launch FB ads without proper tracking and then are surprised when they have no idea what it did for them. FB tends to group everything under the sun as "engagement" and "conversions", so really digging in and understanding those settings is key.
For example, 1-day view-through credit by default is probably a bad idea for many advertisers, particularly when you have no clue what the quality of a view-through is, and what they are worth to you. They are VERY different in terms of value, but FB wants to give them 100% credit with their rules within 1-day. That's simply not how most savvy people approach attribution.
Google Analytics offers some great basic attribution tools out of the box that let you experiment and compare different static models, or create your own static model. Ultimately static models themselves have inherent limitations because attribution is a much more dynamic thing that exists at the individual user path level, but it is a great start.
So many companies launch FB ads without proper tracking and then are surprised when they have no idea what it did for them. FB tends to group everything under the sun as "engagement" and "conversions", so really digging in and understanding those settings is key.
For example, 1-day view-through credit by default is probably a bad idea for many advertisers, particularly when you have no clue what the quality of a view-through is, and what they are worth to you. They are VERY different in terms of value, but FB wants to give them 100% credit with their rules within 1-day. That's simply not how most savvy people approach attribution.
Google Analytics offers some great basic attribution tools out of the box that let you experiment and compare different static models, or create your own static model. Ultimately static models themselves have inherent limitations because attribution is a much more dynamic thing that exists at the individual user path level, but it is a great start.
What are the best attribution tools for mobile (specifically iOS) campaigns?
I'd echo what Jamie said. You can get some visibility on the Android side of things since Google can stitch that together easily. iOS is trickier. TBH I haven't done anything with that specifically yet, so not sure how great the solutions are out there. I'm sure you could do something with statistical modeling, just like anything else, but you need lots of data for that (and a skilled statistician or firm who does this).
Mobile install attribution is an entirely separate thing from marketing attribution. As far as I am aware all the mobile attribution providers are just attributing the last click that led to the install rather than giving partial credit to each ad interaction prior to conversion. The list of common providers there is generally: AppsFlyer, Kochava, Apsalar, Adjust, and Tune (doesn't work with Facebook).
Hi Jamie, I work at AppsFlyer. Indeed the attribution standard is by giving credit to the last click. However, our dashboard provides more insights by providing multi-touch attribution data, meaning assisted installs.
Hey Michael!
I tried to hit on this point in #5 at the end when talking about setting up attribution parameters. We tell most people early stage to go with 1 day click or 7 day click (depending on the business).
If you get the chance to check out Interstate and compare it to GA's attribution tools for static attribution modeling I think you'll be impressed. We have a very generous free plan now (http://interstateanalytics.com/pricing). We should have algorithmic attribution as a premium feature in Q3.
Based on early testing the amount of data required to make algorithmic work is quite high. I'm still not convinced its actually a better solution for companies spending less than maybe $5m/yr. Would love your thoughts on that.
I tried to hit on this point in #5 at the end when talking about setting up attribution parameters. We tell most people early stage to go with 1 day click or 7 day click (depending on the business).
If you get the chance to check out Interstate and compare it to GA's attribution tools for static attribution modeling I think you'll be impressed. We have a very generous free plan now (http://interstateanalytics.com/pricing). We should have algorithmic attribution as a premium feature in Q3.
Based on early testing the amount of data required to make algorithmic work is quite high. I'm still not convinced its actually a better solution for companies spending less than maybe $5m/yr. Would love your thoughts on that.
Feel free to hit me up in Q3 once you have algorithmic attribution in place. I'm actively looking for attribution vendors this year that can tackle the dynamic piece of it.
However my one concern from the limited info on your site is that you only focus on paid media data and don't integrate with other channels. I might be missing something though--like I said, limited info.
On that note, a bit of unsolicited feedback on your site...
I'd love it if you put way more info about your feature set and screenshots/videos on your site without an email gate or anything like that. The candid reality is that while I'm definitely your target audience, I typically have no time to field sales calls without knowing a bit more up front to prescreen. Your marketing site leaves some giant question marks for me as to whether you are even a potential fit, and I don't want to give up my email address and be added to yet another drip list just to find out. I certainly wouldn't just try you out to see if you met my needs because any analytics platform takes some level of setup and infrastructure work that I frankly don't have time for unelss I've determined you're a good fit.
If you have something cool, figure out how to show it to me in as straightforward a way as possible, and don't get it. If I'm interested I'll reach out, if not, you probably wouldn't have sold me anyway.
Also, given how important data and such is to us, your Privacy Policy didn't really do enough to address ownership of our analytics data vs. just our info as users of the product.
However my one concern from the limited info on your site is that you only focus on paid media data and don't integrate with other channels. I might be missing something though--like I said, limited info.
On that note, a bit of unsolicited feedback on your site...
I'd love it if you put way more info about your feature set and screenshots/videos on your site without an email gate or anything like that. The candid reality is that while I'm definitely your target audience, I typically have no time to field sales calls without knowing a bit more up front to prescreen. Your marketing site leaves some giant question marks for me as to whether you are even a potential fit, and I don't want to give up my email address and be added to yet another drip list just to find out. I certainly wouldn't just try you out to see if you met my needs because any analytics platform takes some level of setup and infrastructure work that I frankly don't have time for unelss I've determined you're a good fit.
If you have something cool, figure out how to show it to me in as straightforward a way as possible, and don't get it. If I'm interested I'll reach out, if not, you probably wouldn't have sold me anyway.
Also, given how important data and such is to us, your Privacy Policy didn't really do enough to address ownership of our analytics data vs. just our info as users of the product.
Curious what you mean by other channels, are you talking about email, etc or do you mean integrating data on organic activity as well? (we do both but mostly via just reading UTM tags/referrers).
I do want to build a more comprehensive marketing site, we just have been time constrained on all sides and building out features has taken priority :)
Thanks for the feedback re: privacy policy, etc.
I do want to build a more comprehensive marketing site, we just have been time constrained on all sides and building out features has taken priority :)
Thanks for the feedback re: privacy policy, etc.
Yep, I meant any other trackable channel. For example, DCM right now tracks ad data, but it also seamlessly integrates with GA (and GA Premium). From your site it appeared you just did the ad tracking bit, so I'd want to know how you're baking in other channels (organic, affiliate, email, etc.) and it sounds like for the most part you rely on explicit tagging (which unfortunately doesn't always do it for organic and referral).
Beyond that, until you come out with an algorithmic attribution solution, you really should say more on your site about how you differentiate from GA. Right now I can upload all my cost data, etc. from other engines with no problem, so GA in theory does all of this, and has attribution tools.
Beyond that, until you come out with an algorithmic attribution solution, you really should say more on your site about how you differentiate from GA. Right now I can upload all my cost data, etc. from other engines with no problem, so GA in theory does all of this, and has attribution tools.
Interstate already tracks non-tagged data in the same way GA does.
The GA Model Comparison Tool is great for the basics but I'd hardly call it a comprehensive attribution solution for static models. Even if you import spend data (which you have to do manually), you can't get any sort of understanding of what payback periods actually were, just present day ROAS and how it differs by model. You also can't see revenue, spend, and # of attributed conversion in the same report, filtering is a pain, etc etc. It does the basics but not much else.
The GA Model Comparison Tool is great for the basics but I'd hardly call it a comprehensive attribution solution for static models. Even if you import spend data (which you have to do manually), you can't get any sort of understanding of what payback periods actually were, just present day ROAS and how it differs by model. You also can't see revenue, spend, and # of attributed conversion in the same report, filtering is a pain, etc etc. It does the basics but not much else.
Surprised that the article doesn't mention Website Custom Audiences (WCA in short), which for many startups might be much easier to collect compared to email addresses.
With WCAs you can build audiences based on people who visited your site (or say, a specific URL like a blog post or a thank-you page). You can then use these both for retargeting (e.g. someone visited your site but didn't sign up yet), or build lookalike audiences out of website visitors and/or signups. Imagine that someone read an amazing blog post describing some key features or use cases of your product? Why not create a WCA and then craft a creative addressing that specific audience segment.
Another benefit of WCAs and lookalikes based on those is, that they're updated automatically, whereas for email-based custom audiences you'll either have to manually upload new signups to FB, or set up some custom automation to do that for you.
With WCAs you can build audiences based on people who visited your site (or say, a specific URL like a blog post or a thank-you page). You can then use these both for retargeting (e.g. someone visited your site but didn't sign up yet), or build lookalike audiences out of website visitors and/or signups. Imagine that someone read an amazing blog post describing some key features or use cases of your product? Why not create a WCA and then craft a creative addressing that specific audience segment.
Another benefit of WCAs and lookalikes based on those is, that they're updated automatically, whereas for email-based custom audiences you'll either have to manually upload new signups to FB, or set up some custom automation to do that for you.
Facebook is an effective acquisition channel for many verticals especially for Gaming and E-commerce companies. For Mobile gaming, activation campaigns benefit from Facebook's mobile reach, targeting and analytics. For e-commerce, Dynamic Product Ads will generate the highest ROI among most paid marketing activities (albeit at limited volume).
There is one huge catch, though: measurement and attribution.
Yes, setting view and click attribution windows on the platform is a given and necessary, but this is only partly effective, unless Facebook is the only paid marketing channel you're using.
If you're any running media on other networks or channels, then you need to measure the their interactions and influence on the customer journey in order to arrive at the incremental value of each channel and an attributed CAC (customer acquisition cost).
Most attribution partners will allow you to do this and play nicely with all of the networks/partners/channels, with one exception: Facebook.
Facebook is a 'walled garden' in that it does not allow third-party impression tracking, unless you're using its attribution product, Atlas.
This means you are unable to effectively value and weigh the effect of Facebook impressions, measure frequency and overlap across channels, conduct accurate path analyses, and understand the incremental value of Facebook.
That said their ad products are sophisticated, best in social IMO, and their roadmap is very promising.
There is one huge catch, though: measurement and attribution.
Yes, setting view and click attribution windows on the platform is a given and necessary, but this is only partly effective, unless Facebook is the only paid marketing channel you're using.
If you're any running media on other networks or channels, then you need to measure the their interactions and influence on the customer journey in order to arrive at the incremental value of each channel and an attributed CAC (customer acquisition cost).
Most attribution partners will allow you to do this and play nicely with all of the networks/partners/channels, with one exception: Facebook.
Facebook is a 'walled garden' in that it does not allow third-party impression tracking, unless you're using its attribution product, Atlas.
This means you are unable to effectively value and weigh the effect of Facebook impressions, measure frequency and overlap across channels, conduct accurate path analyses, and understand the incremental value of Facebook.
That said their ad products are sophisticated, best in social IMO, and their roadmap is very promising.
Facebook allows a few other third party tools for measurement/attribution analysis. Ref https://facebookmarketingpartners.com/marketing-partners/#ad....
No, they cannot provide measurement and attribution for campaigns that use Website Custom Audiences or Lookalike audiences, which are often the two largest, most valuable segments of any campaign. Atlas is the only provider that can.
Also, this list points to mobile measurement companies.
Also, this list points to mobile measurement companies.
I've been advised that advertising enterprise products (for US market) on Facebook are a bad use of resources. Then we looked at Linked In, but their pricing and targeting was disappointing.
What's the best platform to advertise enterprise software on? I think IP level targeting is a good strategy, but curious to hear about other ideas.
What's the best platform to advertise enterprise software on? I think IP level targeting is a good strategy, but curious to hear about other ideas.
Is your target audience on FB? That's question number one. Number two is whether there's anything relevant you could put in front of them that would stick or get some engagement.
Also, for most paid social efforts you need the ability to look at attribution across channels. Oftentimes paid social plays more of an awareness role that makes it look like it doesn't perform worth a damn if you are comparing it on a last click basis to say, paid search.
Also, for most paid social efforts you need the ability to look at attribution across channels. Oftentimes paid social plays more of an awareness role that makes it look like it doesn't perform worth a damn if you are comparing it on a last click basis to say, paid search.
Your best option is Account Based Marketing. Look into Demandbase.
Facebook does have a brand new ad unit for mobile click to call, and I have seen success with b2b with Facebook ads.
The key is to get your ad unit as native as possible to the atomic consumption unit of the site. So, since facebook is used to share content, inspirational quotes, etc... maybe try promoting a whitepaper, a video or some other strategies.
Bottom line, Facebook right now has by far the most sophisticated targeting of any ad tech that is publicly available (Short of some private DMPs) and their lookalike audience can match over 2000 variables about your audience. There is no better way to find the right people.
Also consider that FB is mostly a mobile product now and that more video is consumed on FB than on Youtube. and they have a broad reach with their audience network.
Facebook does have a brand new ad unit for mobile click to call, and I have seen success with b2b with Facebook ads.
The key is to get your ad unit as native as possible to the atomic consumption unit of the site. So, since facebook is used to share content, inspirational quotes, etc... maybe try promoting a whitepaper, a video or some other strategies.
Bottom line, Facebook right now has by far the most sophisticated targeting of any ad tech that is publicly available (Short of some private DMPs) and their lookalike audience can match over 2000 variables about your audience. There is no better way to find the right people.
Also consider that FB is mostly a mobile product now and that more video is consumed on FB than on Youtube. and they have a broad reach with their audience network.
Facebook have recently launched Lead Ads [0] which could make them work for an entreprise product campaign if you can find the correct targeting. Worth trying it out.
[0] https://www.facebook.com/business/news/lead-ads-launch
[0] https://www.facebook.com/business/news/lead-ads-launch
Don't advertise, do lead-generation & sales.
Read this: http://www.amazon.com/Predictable-Revenue-Business-Practices...
Read this: http://www.amazon.com/Predictable-Revenue-Business-Practices...
Advertising is a part of lead generation. Content marketing with paid distribution is a valid strategy that is in use by a ton of SaaS companies today operating on the Predictable Revenue model.
yes, that's probably the way to go for enterprise SaaS businesses. I should have been more specific, though.
For brand advertising for Enterprise Saas, not lead generation, what are some good strategies?
For brand advertising for Enterprise Saas, not lead generation, what are some good strategies?
One thing I would add is that you don't need a large list to make good lookalikes off of - some of the best lookalikes were generated from only 5,000 emails.
Lookalikes almost always outperform interest targeting.
Lookalikes almost always outperform interest targeting.
Past some point yes. I've seen a list of 5k work. I've also seen a list of 5k fail miserably and get outperformed by interest audiences. I don't think its accurate to say that interest audiences never outperform lookalikes. At some small scales they do.
Also there is a huge difference between a list of 5k emails that signed up and 5k emails that purchased in terms of performance.
Also there is a huge difference between a list of 5k emails that signed up and 5k emails that purchased in terms of performance.
You can also combine lookalikes with interest targeting -- once some audience gets saturated, mix a bunch of your past performing segments like a lookalike + interests.
Agreed. I was probably a bit too general.
Wow. I didn't know you could upload a list of your customer's email addresses to Facebook and have it build an audience of similar users. This seems highly unethical -- who knows what FB will do with your customer's email addresses.
Edit: I also wonder what would happen if I uploaded just my own email address? Would it find people very similar to me?
Edit: I also wonder what would happen if I uploaded just my own email address? Would it find people very similar to me?
The data is hashed locally to get around personally identifiable information (PII) issues: https://www.facebook.com/business/help/112061095610075
You can't upload just one email address (they changed the minimum number a long time ago for privacy reasons). There is a minimum number to create a custom audience of email addresses and then you create a lookalike audience that is 1MM+ people. I don't see how that's a violation of privacy.
It's probably the most effective thing Facebook has done for targeting as I use almost exclusively custom audience and lookalike targeting.
It's probably the most effective thing Facebook has done for targeting as I use almost exclusively custom audience and lookalike targeting.
This is now a universal feature offered by Google, Bing, Yahoo Gemini, Twitter, Linkedin and I'm sure others. It is no different than cookies for retargeting. leveraging first party data is by far the biggest opportunity in advertising and this trend is here to stay.
No one is using that data except the advertiser who uploads it.
Great article
Turns out that the first thing you need to do is figure out if FB is the right channel for you. I found out that on FB, anyone will download anything that looks interesting and you can optimize your CPI fairly easily. But if you count on people spending money via in app purchases, the typical rules (1%-5% of active users) don't always apply for apps of different genres.