Some Personal News

43. Unlocking Data Clean Rooms + a Deep Dive into TV

Jun 11, 2023

Happy Sunday. To those I spoke to this week or spent time with in-person - thank you. A very warm welcome to all the new subscribers. I’m thrilled to have you as readers and truly appreciate your feedback and support.

Let’s dig in.

Get Deeper into Donor Behavior

The term “Clean Room” on its own is defined as a “secure environment where several parameters are constantly monitored and controlled.” Nothing should leave the Clean Room, but you can do some magic inside it.

Data Clean Rooms are the same. 1^st party donor data must be a protected asset.

It wasn’t long ago that Google were zeroing out DoubleClick IDs to limit marketers’ ability to stitch individual ad impressions across devices, websites, and time into a complete user profile.

That measure was an immediate but temporary overkill to prevent legal exposure while the industry figured out what to do next without 3^rd party cookies. Data Clean Rooms emerged as an answer employed by individual publishers and larger media buying platforms. To illustrate how it works, we need to imagine a world of advertising without 3^rd party cookies with no substitutes available.

Take New York Times (NYT) as an example. They have lots of information about their readers – personal details, payment methods, readership history, some modeled behavior, and the history of their interactions with ads on NYT.com:

On the other hand, there are advertisers – such as P&G – who also have access to lots of data about their consumers, some of whom potentially overlap with the NYT readership base:

In our scenario, NYT wants to work with P&G to sell ads and P&G intends to spend its marketing budget responsibly, find more prospects, and get in front of them. But P&G doesn’t know their consumers are avid readers of NYT.com, making NYT a suitable medium for P&G to spend on – the 3^rd party cookies that would enable this insight are gone. P&G is forced to rely on content targeting instead of audience targeting without distinguishing between NYT and other media outlets with similar content. This, in turn, hurts NYT’s ability to sell personalized ad solutions to P&G.

Without 3^rd party cookies enabling advertisers to target a specific audience regardless of where they are, advertisers become blind in their targeting. This blindness hurts advertising performance – and publishers’ ability to develop deeper relationships with advertisers to increase their own profitability.

It’s in NYT’s interest to put out a tool that allows advertisers to “match” their current consumer data to the NYT readership base. This is where the Clean Rooms come into play. NYT develops an environment where P&G can upload their existing consumer file and see how many of them are NYT readers, what content they consume, how frequently they visit NYT.com, and other information.

A close-up of a paper

Description automatically generated with low confidence

P&G doesn’t get access to data on all the NYT readers – only the ones it already has on file. Moreover, no sensitive information leaves the Clean Room – P&G only gets the aggregated insights back. But by doing so with NYT and 100 other publishers, P&G can now say where the match rate with their current consumers is higher – therefore assuming that NYT has more readers who look like prospects for P&G.

Clean Rooms provided by programmatic platforms with access to multiple publishers – Google, Amazon, Quantcast – work similarly. Instead of publisher-level overlap, they enable advertisers to know which media platform has proprietary targeting options that match their consumers better.

Data Clean Rooms have two primary applications for marketing teams:

Analytics – by uploading a subset of the donor file to a select Clean Room, marketing teams can learn which content those donors consume and use that for creative development or more accurate content targeting.
Media Planning – by uploading current donor segments to various Clean Rooms, marketing teams can understand which platforms or publishers overlap most with the organization’s donors.

The latter use case is a game changer, allowing you to do better, publisher-level “look-a-likes” without 3^rd party cookies. But engaging with multiple Clean Rooms is very resource-consuming. So, to enable advertisers to simultaneously get access to numerous publishers and not spend months building the infrastructure, “aggregation” of Clean Rooms is becoming a thing now. One of the first companies doing it was Habu – check them out; they have a lot of good, deeper content on this topic.

Data Clean Rooms are becoming a new major source of data enrichment to get deeper into donor behavior. Working with them currently requires significant data engineering and BI capabilities that not everybody has access to – but if you do, crack on and make magic!

Deep Dive: Exploring TV and CTV

TV and CTV is one of the most under-discussed marketing channels in non-profit marketing, and ecommerce in general. I don’t see many articles or threads about this topic, but I know there are a few people in my network who are responsible for a lot of success using it.

Before wading into TV, consider the following bullets:

You have battle-tested donor acquisition funnels that you’ve tested and optimized to be the best they possibly can be.
You have the ability to test TV with a solid $150k budget that you’re OK to lose.

And allow me one disclaimer: I don’t know how much of this applies to brand awareness TV campaigns. I’m not writing this post through that lens, so fellow brand marketers please share with me your perspective.

Here’s what we’re going to cover in this Deep Dive:

Why TV Ads
Pricing
TV vs CTV vs OTT Inventory
Measurement
Creative

Why TV Ads?

1. It’s cheaper to test than you might think. Remnant ads can be $1 to $3 CPM to get on a prime network like ESPN or CNBC. Meanwhile on Connected TV (CTV), you’re more likely to pay $10-30 CPM but you’re also targeting the user, not the content.

2. Increase in search traffic and awareness. If your creative does a phenomenal job explaining who you are and what the problem is that you solve for the world, you’ll see a direct correlation to branded search traffic and inbound interest. By getting someone to even type into Google “Your Organization Name + TV” and then seeing the TV offer linked in the sponsored Google ad up top, you’re WINNING.

This is where it’s extremely important to have a proper funnel built out that is battle-tested. I’ve built a few TV-focused landing page funnels in my time and for this to work well, have found that teams need to:

Ensure the creative on the landing page is similar to that of the TV commercial.
Ensure that the landing page does a good job of collecting first-party data (1P data), as quickly as possible. This might even mean running a "giveaway to increase the opt-in rate of the 1P data, but remember - it’s still a pre-qualified audience if they’re on this LP from the TV commercial (they’re showing high intent here).
Have a bullet-proof WHY: “why your Org and why now”.
Ensure you communicate the credibility of your Org, the safety of donor information (those credit card secured logos and icons), what the Org is, the problem you’re solving, and how soon they can help it if they donate right away.

Overall, you have to treat TV traffic like it’s very upper top of funnel traffic that can leave at any time, so there’s no time to BS on the landing page. Be fully in service to the prospective donor coming to your Org’s landing page, and make sure you deliver value right away.

3. Increased performance on other existing ad channels. A lot of Org’s hit a point of diminishing returns on platforms like Meta or Google. There’s a ceiling for when they spend their budgets and the CPA just balloons to a point where it doesn’t make real sense to continue scaling the account budgets. Running TV expands your brand awareness so much, allowing for you to increase the ceiling on that point of diminishing returns for your digital ads. With some Org’s, I’ve seen Facebook budgets increase by more than 3-4x within 18 months of turning on TV ads.

Additionally, for the months and quarters after you run TV it’s very common for Org’s to see their CPA decrease on their other existing ad channels. There’s a “halo effect” that brings second order efficiencies to your existing digital ad spend on Meta/Google/TikTok etc.

Pricing

Transparently this is a bigger pill to swallow than what you may be used to, if you haven’t previously gone outside of the digital channel ecosystem. It’s not an easy thing to stand up. Just testing something once doesn’t work, you have to be in-market for a little bit to really understand the benefits you’ll reap from the channel.

When testing any new channel, there’s obviously a high chance you’ll be able to make it work, but you should always think about test budgets as something that you’d be ok with losing completely. It’s like a gamble. Hopefully, with the right strategy, tactics, and smarts you’ll be able to make a channel work for you within your test budget, but worse-case scenario, think of it like an investment.

Out of the $150,000 I mentioned above, here’s how you should think of that breakdown:

$80,000 - $100,000 is reserved for media buying.
- This allows you to properly test what works and what doesn’t work with enough room to test different placements, networks, and types of TV media to buy.
$10,000 - $30,000 is reserved for creative, though you can go up to $100,000 for a really high-quality creative piece.
- This budget can scale depending on the production value you believe you need to have and allows you to produce, edit, and test enough angles and hooks.
$20,000 - $50,000 for an external partner to help you navigate all of this.
- With any new channel, it’s smart to hire a channel sniper, which is what I like to call industry people who know the channel’s industry, people, practices, tips, tricks, strategies, and insider knowledge inside and out. The same way you know your own Org’s mission and value-add inside and out, that’s how some people know media channels.

If you don’t have ~$150,000 to set aside, that’s fine - just wait. There’s no point in doing this without everything above in place. You’ll run into the same problem that every company runs into of “We took the best creative on this channel, and ran it on the new channel and it didn’t work, therefore this channel doesn’t work and it’s a bad investment.” And then they wonder why they’re 15 years behind the leading brands in their own categories. Don’t test with a lazy approach - you need fresh strategy, fresh creative, funnels that match the donor experience, etc.

Lastly, the timeline for this $150k investment is about 8-12 weeks. Within that timeline, you can do everything I’m writing about in this post.

TV vs CTV vs OTT Inventory

For the most part, performance-marketing-focused Org’s buy media on CTV and linear TV. You can absolutely buy on OTT platforms too but for many the CPMs don’t make sense and there isn’t a proportionally higher return on the investment to justify the CPMs.

For non-profit Org’s (and DTC brands in general), remnant TV ads on linear are the perfect place to begin your TV strategy. They're often significantly cheaper than prime time slots - the trade-off is that these slots might reach fewer people, and it might be at odd times. That said, “remnant” doesn’t just mean overnight or low viewership inventory. You can buy remnant inventory in any rotation (prime, weekend, etc) on any network, at heavy discounts. This includes buying up unsold ad inventory at $1 to $3 CPMs on huge networks like ESPN, HGTV, CNBC or smaller networks like The Outdoor Channel, True Crime Network or The Country Music Channel.

With this approach, the breadth of reach, and ability to directionally target your donor base (you know the type of person who’s watching ESPN versus who’s watching HGTV) is where the most value lies.

Once you find your footing with TV and optimize the creative, offer, and funnel for performance, then it makes sense to layer in streaming (OTT) platforms, first through low-CPM publishers (Paramount+, Viacom, Tubi and Pluto), which are generally sub-$10.

OTT content is always delivered through a CTV device, meaning you have to be able to use a smart TV, an Apple TV, or something like a Roku to access Paramount+ in the first place. The OTT player’s ability to provide an ad exchange, and the CTV player’s ability to provide data and insights into the viewer is what makes this a powerful combination.

Budget-wise, your OTT/CTV budget is generally 20-30% of the linear TV budget. It’s not as strong for finding new donors, but it does really well with retargeting, and as you get ramped up on linear, the prospecting side of streaming TV ads works more effectively. The longer you’re live on TV, the more efficient you’ll feel as the conversion cycles will feel quicker.

If you can’t afford to start with linear TV, then the second best place to start is by using OTT/CTV placements to just retarget website visitors with ads for your Org. Your commercials will air on CTV devices of prospective donors who’ve been to your site and not donated, and because it’s a programmatic buy, you can scale these budgets up and down with your own Org’s seasonality. Plus, with streaming, a 15-second and 30-second commercial cost the same amount, so you get some more room to demonstrate why you’re great.

Measurement & Attribution

I reckon the best approach for measurement and attribution is two-fold.

1. Donor Post-Donation Surveys. Surveys are the absolute best way to measure lower-funnel performance. While your GA data might argue that this person came from a branded search ad, the post-donation survey will tell you they were prompted by the TV ad that aired on a specific channel.

2. Spike Measurement Analysis. You can do this directly inside the platform where you buy the media or your attribution platform, and use this data in conjunction with surveys to understand which individual ad creative, network, and rotation contributed to the donations generated from TV.

Creative

Fundamentally with TV creative, you want to tell a very clear, concise, and easy-to-understand story, including a demonstration of the problem you’re solving, the ingredients that make up your mission-proposition-solution, how the solution works, and the benefits of the solution to the recipient (at the end, focusing on benefits specifically, not value props).

Unless you’re Nike, I strongly encourage you not to run abstract, motivational creative especially without a crystal clear ask. Unlike with paid social ads, concise storytelling is more valuable than abrupt jolts and quick cuts.

As you produce your creative, you want to focus on generating 2-3 concepts that you can test with 4-6 angles or hooks in the beginning to A/B test. In your planning of the creative, you should overcompensate for what you think you’ll need. Get all the extra shots, cuts, takes, and angles you think you’d ever want to test, within reason. The beauty of TV creative, is once something gets tested and starts working, you can have a piece of creative efficiently run for 3 years in a row, with no change to the performance.

That’s it for today. Thanks for sticking with me in this longer edition. Does that answer most of your questions about how to approach TV and think about Data Clean Rooms? What else would you like to know? If you have more questions, just reply to this email and I’ll do my best to review and respond.

Now onto the fun stuff »