Some Personal News

#11 Your Custom Bidding Playbook, Becoming Adaptive and a Channel to Explore

Oct 16, 2022

Happy Sunday. A very warm welcome to all the new subscribers. I’m thrilled to have you as readers and truly appreciate your feedback and support.

In this edition of Some Personal News:

Deep dive: exploring custom bidding for enhanced performance + a playbook
How taking an adaptive approach to your channels/platforms will deliver a superior donor experience
A third marketing channel outside of Facebook and Instagram that Non-Profits could be exploring to acquire new donors

If you don’t find this content valuable, you can permanently unsubscribe at the bottom of this email. If you like it, subscribe and don’t miss the weekly post:

Let’s dig in!

News to Peruse

Google released two updates for Display and Video 360:
- 1) Optimized targeting to help increase reach by combining machine learning with your campaign settings.
  2) Exchange Provided Identifiers (EPID) to help Display and Video 360 exchange tracking data with first-party publishers to improve ad quality.
Twitter's trying another way to enhance community engagement creating dedicated audio Spaces within their groups function
Introducing “Handles: A new way to identify your YouTube content
At TikTok World Summit the brand showcased a feature where brands only pay if their ad is viewed or interacted with within 6 seconds

Enhance Performance with Custom Bidding

We always cover “Performance” in Some Personal News and in last week’s edition I mentioned Custom Bidding, and included this diagram:

Timeline

Description automatically generated

Custom Bidding is one of the most widely-discussed topics in the context of managing SEM, Paid Social, and Programmatic Display campaigns, and I was delighted to receive some fantastic questions in response to last week’s edition. Thank you.

First, let’s set it straight: Custom Bidding is usually overused. Under-understood and overused – leading to most accounts achieving less than half of their potential, wasting money, and deteriorating trust in the algorithms. And therein lies the opportunity…

A good friend of mine and a long-time SEM marketer in the travel vertical, has a distinct philosophy of never using any bidding automation in Google Ads. Not even Enhanced CPC. She’s managed upwards of $250M over her career – all on manual CPC bids. Why so? She told me the following: “Automation makes you vulnerable. You give up your control and revenue to Google – they decide who to give a likely conversion to, you or your competitor. And they decide based on their KPI, not yours – the Ad Spend. Who spends the most gets the most.”

While that’s a pretty extreme example, here’s a more relatable one. I interviewed a Facebook Paid Social manager last quarter who, when asked about their optimization approach, told me the following: “You set the geography to nationwide US, upload the image assets, and let Facebook’s algorithm figure it out.” The interview was over right there.

In medio stat veritas. My 5th grade Latin teacher would be proud. The truth is somewhere in the middle. Automation is a good thing – but only as long as it works FOR you, not INSTEAD of you. You keep and enhance the insights, and Tech does the work you can’t, like setting bids to $1.01084 instead of $1…

Over the years of implementing automation in Google Ads, Facebook, DV360 (DBM), TTD (The Trade Desk) and many other tools, I came up with a playbook - I’ve listed the 5 core elements below.

But first there are two considerations:

Choose the right system. Google Ads was created for CPC marketing. Facebook and Programmatic platforms were designed to buy on CPM. That difference in nature determines which automated algorithms work better – Google Ads is better off with conversions set up as click-through only.

Use the latest tech for testing. Google Ads is Google’s “beta testing platform,” meaning it gets the “latest and greatest” untested versions of the machine learning algorithms. Know this – Google Ads is currently using the Google “Brain AI” v4, while DV360 is on v2. DV360 is likely better enabled to develop your custom algorithms, but out of the box Google Ads can drive better results.

The Playbook:

Always start with manual bids. This helps you build the “muscle strength” of the campaigns, which comes with the proper structure. You are always better off starting with the basics for the first month or two. It will force you to do the basics right – break the campaigns into audience segments, analyze device and browser performance, and derive other crucial insights. Starting manually will also help you set the reporting structure so later you’ll be able to judge whether automated bids give you the same volume of insights necessary to understand what works.
Start with the lowest level of automation. In Google Ads or Facebook, that’ll be Enhanced CPC. In Programmatic platforms, “Minimize CPC” type of bids. These low-level automation techniques derive the key value from the machine – Accuracy.
Work your way one by one. Don’t rush from CPC bidding to Target ROAS – chances are, campaigns will fail immediately. You either won’t be able to spend or will lose any sight of what works. Start with the “Minimize CPA” tactic. Once that one hits the limit – determined by three consecutive days of the CPA changing by no more than 1% a day – shift to Target CPA, setting it just above the stalling point. Then reduce the target CPA gradually every day until you hit the limit and cannot spend the assigned budget anymore. That is likely the right time to shift to the Minimize ROAS and do the same play there.
Break tactics into funnel steps. You’ll start noticing behavioral differences for different campaigns and audience segments as you go through the algorithm in the first 3 steps. Lean into that instead of neglecting it. The higher your tactic is in the funnel, the more likely it is to work worse with Target CPA – leave aside the Target ROAS. Bidding algorithms are not created equal for a reason and you’re better off using the “Engaged Visit” or “Pillar Page View” instead of a donation as the Action to target.
Consider going beyond the basics. Once Target ROAS has exhausted itself, the time has come to consider more advanced, custom algorithms. It is too much to cover in this newsletter, but that is when you want to connect your CRM to the cloud and develop your custom bidding algorithms based on the Donor’s Lifetime Value. You might need outside help unless you have an in-house data engineering team.

Remember the insights. Automated bidding works – but only if you put it within the constraints of the campaign structure that you can manage manually as well. That way you can still run all the reports you need and present to the exec team a meaningful view of what works and what doesn’t.

Boston Consulting Group seems to agree – Tech with human input works the best:

Become Adaptive to Each Platform

The last decade of non-profit advertising (specifically 2013-2018) was rather nice. Facebook created a monstrous ad product that allowed us to reach virtually anyone. Within 30 minutes you could have a pixel on your site, an ad account set-up, and traffic would start arriving on your site at a very fair priced CPM.

When it came to reporting we could look right at the Ads Manager dashboard and you'd see that for every dollar you spent, there's a return. Everything just worked!!

Today, it's obviously much different. Even just getting pixels, domain verifications, an ad account that doesn't get disabled within 24 hours, or the way you verify your non-profit to run ads are all factors that make today's digital marketing more complex.

People love Facebook ads because in theory it's easy to set up, the targeting capabilities are incredible, the list-matching is incomparable, and the best part is, your CFO can see the return right there. You don't have to run it through a data scientist or any fancy software - it's right there, under the ROAS (return on ad spend) column.

When you look at a platform adjacent to Facebook it feels inferior. Let's take TikTok that we covered in Edition 9, for example, which in my opinion offers one of the best opportunities for reach right now.

“The traffic on TikTok isn't that great, it only stays for 10 seconds."
"The audience is too young and they probably don't have any money to donate."
"We don't get much engagement on the ads we run on TikTok so we decided to scrap it."

This isn’t just specific to TikTok. The problem is that most non-profits are running ads on other platforms in the same way they run ads on Instagram.

You have to become adaptive to the way these platforms push out content, the culture within each platform (understanding the trends, the memes, the comment sections, the way the swiping works, etc), and then create experiences around that.

If you run a fast-cutting Instagram ad designed for a “sound-off” experience that goes to your homepage, and expect it to work on almost any other channel, you'll never find success.

Ad platforms provide the opportunity to start the conversation with a new donor, but they shouldn't be the only thing in your donor journey. You need to make sure there’s a proper journey after the social ad.

It’s important to focus on what the context of the platform is, and that you’re doing a good job of being present. Bit like going into a new city or country - you need to be relevant there, not where you came from. Take it from an Englishman! With that said, take a look at Outbrain/Taboola below:

Exploring a New Channel (Part 3):

Have you ever scrolled to the bottom of an article and seen the section titled "You May Also Like..."? That's Outbrain. Taboola is just like it, and they both do pretty much the same thing.

There are 3 types of site experiences I've seen work incredibly well with these native content ad units:

Really lengthy landing pages with inserts of editorial content
Listicles/articles/comparicles (listicles comparing 2+ programs/brands)
Pointing to really good earned media content, in which case after the click, Taboola or Outbrain wait for the prospective donor to hit the organization’s site

Often the page is insanely long and can look pretty sketchy. Somehow, this works and drives a great CPA for the org (cost per action).

I know some ecomm/DTC players have used it on Facebook a lot, but unsure if they ran it on native ad networks - either way, it would be great content to live there.

With these content networks, your goal is to launch, understand the types of publishers that work, narrow down to specific publishers, and then ramp up spend. You should be looking for new donors here with short conversion windows because you lean on content for the education of the donor.

Interesting Reads This Week:

A breakthrough in AI utility, this time with voice - Joe Rogan interviews Steve Jobs. All created by AI. An ethical debate has ensued.
How did Uber apply behavioral science at scale? By going the extra mile on data.
How to make your non-profit boom in a recession? Be irrational.
You might be doing a lot more than your Apple Watch is giving you credit for.
Why we don’t need to put an upper bound on productivity growth to deliver sustainability