Hello experiMENTAList, it’s Sundar 👋

I’m a former Head of Marketing Science at Uber where I optimized $1Bn+ in spend across Brand, Performance, and Lifecycle.

Now, I’m sharing the same playbooks top Consumer Techs use to prove, optimize, and scale Marketing ROI with you!

Thank you for being a valuable part of this growing community of 2.6K+ marketing leaders from Uber, DoorDash, Google, Spotify, and many more.

Now let’s step into the lab!

PS: If you know a colleague or friend that might benefit from this newsletter, send them here.

Whenever someone responds to my welcome email, I ask them if there’s anything within Consumer Tech Marketing (Strategy or Science) they’re interested in reading more about.

Here was a recent ask:

And if the question is in my wheelhouse then I write an article about it (which is why I encourage all of you to tell me what you’d like to see!).

So today we’re going to discuss what you need to do to scale Meta Ads.

We’ll cover:

Let’s go scale!

Before, we dive into how to scale on Meta, there are 2 fundamental questions we need to answer.

I’ve scoured the internet and the general consensus is that “knowing you have a scaling problem” is truly more subjective than you might think, but it boils down to does more budget = more performance.

So, simply, you have a scaling problem when increasing budgets results in a linear increase in CAC or decrease in ROAS.

For example, if you increase budgets by 20% and your CAC goes up by 20% (or ROAS down by 20%) then you effectively do not have the right environment to scale.

This is easy to answer but hard to actually assess (no one said this would be easy).

You need to keep pushing your budget until cost per marginal incremental acquisition = marginal LTV.

That sounds like a lot of words so let me break it down a bit.

The cost per marginal incremental acquisition looks at how much you additionally have to spend to get that last incremental acquisition.

Example

In this example, the cost per marginal incremental acquisition of going from $100 → $200 in spend is $4. Why?

($200 - $100) / (75 - 50) = $4

For an additional $100, you get 25 more incremental acquisitions at a cost per marginal incremental acquisition of $4.

Now, let’s look at the LTV side. At $100 in spend, you generated $500 in LTV. At $200 in spend, you generated $675. This means that the incremental 25 users generated an incremental LTV of $7. Why?

($675 - $500) / (75 - 50) = $7.

In this situation, the incremental users you acquire have a marginal LTV of $7 while your cost per marginal incremental acquisition = $4. Keep scaling baby!

Remember, attributed ≠ incremental so only incrementality tests can help here, but if you’re still early in the scaling journey, you might not have to worry about this yet.

Now let’s look at what else we have to do to scale Meta!

The biggest news in the Meta world is the launch of Meta’s Andromeda which is their AI (what isn’t at Meta) based ad distribution engine. Essentially, it takes their Advantage+ (aka ASC+) campaigns and is now using that engine to power all ads.

The beautiful thing is that it simplifies a lot of complex account and campaign structure challenges that many people had. The consensus is that Andromeda is insanely powerful and that there’s no point fighting Meta’s ability to serve the best ads (much like experts feel about Google too) so luckily you have one thing less on your plate.

So what do you need to focus on instead?

Meta is now doubling down on the power of creatives. Half assed creatives will no longer suffice and the real advantage Marketers have is in building high quality creatives . Yes, quantity matters because you need to test different concepts but the true differentiator is now quality.

So, if you’re hitting a scaling wall, here’s a few things you must do on the creatives side:

Like I mentioned, Andromeda seems to be rewriting the rules of campaign structures and set up as we speak so there’s still a lot of learning left here, but one thing is clear:

If you are using Meta for your acquisition campaigns, then you should be damn sure that the campaigns are primarily driving new user acquisition and not existing users. Targeting existing users is the #1 reason platforms show incredible ROAS making your ads look more incremental than they truly are.

Now, retargeting in itself is not bad but most companies are really bad at accounting for that when calculating LTV or even identifying users as existing users, so a clear way to prevent this is by ensuring you’re not targeting existing users.

However, when you shift from targeting existing → not targeting existing what might happen is that your ROAS will drop BUT your scalability improves because now it forces Meta to go after new audiences and reallocates budget to more incremental users.

As part of this broader “don’t target existing users” tactic, you should also scale budgets up on campaigns / creatives that are driving the most amount of New Users by looking at the breakdown by audience.

Otherwise, there’s doesn’t seem to be other major account structure changes you should use to scale on Meta, but give me feedback if I’m wrong and I’ll update the post.

There is, however, one campaign structure change you should implement if scaling is an issue for you.

A common reason companies hit scaling issues is they’ve tapped out of low hanging fruit. This means that they’re now circling the drains asking Meta to acquire cold customers.

To combat this, introduce middle and upper funnel creatives that create more warm customers. These are campaigns that tend to have a higher frequency using creatives that introduce the brand and the goal is reach (not conversion).

You’ll then have to pair this with your lower funnel conversion campaigns creating that classic “funnel dynamic”.

Here’s an example:

This allows for higher volume of scaling and is an eventual step for all large brands.

Now we’ve spent a lot of time on what to do in Meta but there’s just as much to do outside so here we go!

Talk to any world-class marketer, and they'll tell you that both product and data science are their best friends. That's because no matter how good the marketer is, there is a theoretical ceiling that they'll hit unless they have these partners they can work with.

So, if you’re a marketer, what do you need to ask for from Product?

The biggest barrier to acquisition is not just marketing but also product. And if we dig into it more specifically, it's a broken promise between what marketing says and what product delivers. So, naturally, the quickest way to drive scale is to be able to improve activation rates. This has two fold impacts:

In the lens of how to scale Meta, what we need to do is deliver optimized landing pages that are customized for the creative messaging, allowing users to see a message and a promise, then click on the ad and go to a landing page that reinforces that message and promise, and then eventually convert and have that promise delivered.

For some reason, companies still do not see the connection between ads and a landing page, but it is a SUREFIRE GUARANTEED way to improve CAC and help with scaling.

There are of course other product investments a company can make but they’re bucketed in the broader category “Just make my product better”. Given it’s not specific to Meta, I’ll leave that out here.

All right, baby, now we're getting into the fun stuff. And I hope all the MarTech readers are sitting there giddy.

So if you're an e-com business, the #1 MarTech investment you can make is on fixing your product feed. This means better images, accurate sizing, complete inventory, and a whole host of other fixes that you have to make before you think of doing anything else to scale Meta. The impact on this will be almost instant because it means that Meta is serving more accurate ads and therefore better conversions, which creates a beautiful feedback loop.

Another key investment you need to make is on server-side tracking through Facebook's Conversion API. But it's not just setting up this API (if you haven't already). It's also adding valuable information for Facebook to make better decisions.

For example, don't just share that a conversion happened, but also share the value of that conversion in revenue and in profit or margins if you have that as well.

Finally, another key MarTech investment is an updated exclusion list. This is useful for preventing existing users from being targeted, as I shared before. At first, you might need to do it manually, but the sooner you can automate this list and keep it up-to-date, the better to scale

That's pretty much it on the MarTech investments that are absolutely critical to scaling.

Now you might be surprised that I'm not including anything related to attribution, but frankly I don't believe attribution alone helps you scale at all. So investing more resources in getting attribution correct, no matter what version you use, is actually not productive.

As a data scientist, my whole career I've always been confused by why marketing and data science teams complicate things. In my experience, there are only two sets of analyses that really help you scale.

This is the number one experiment you can use to scale, and it might seem counterintuitive that I'm telling you to turn off spend. The incrementality test has two insane benefits:

When you are able to go to your CEO or finance team and say, "I know our marketing budget was $100, but we spent $80 and got the same results," that is incredibly powerful.

Now, without getting into the political trap of "But I might lose that budget next year," you can use that $20 to go test out new channels or new experiments or pay for something else rather than giving it to Zuckerberg.

Incrementality tests are also the litmus tests of a culture that is receptive to experimentation and being truly data-driven. This is the foundation that is rarely spoken of but always present at every single Top Consumer Tech.

A word of caution: There are two types of incrementality tests. Channel level (aka is this whole channel incremental) vs Spend level (are we incremental at this spend).

More often than not, you should be testing spend levels and not channel level. There are only 2 times Channel level incrementality will show a channel isn’t incremental:

The second analysis I recommend is channel-level funnel and LTV. For the purposes of this exercise, any attribution model is fine. What you want to do is segment your funnel conversions and LTVs by attribution channel.

Example

Fun fact: I'm red-green colorblind, so charts like this are absolute nightmares. But let's look at the line that's all the way at the bottom, which I believe is “Display”. The LTV looks to be about half of the top line.

This analysis simply shows you how much more or less effective channels are and acts as a signal for where you should test incrementality. Using the example above, if I were running an incrementality test, I’d start with the Display channel and see how truly incremental it is.

And that’s it for how you scale on Meta.

Now, depending on the level of scaling you’re at you can pick one or two of these but the more budget you have and the more you scale, you’ll have to action all of these items. The companies spending hundreds of millions do ALL of these from account structure to product investments along side high quality MarTech stacks and data science.

There’s no real short cut to scaling Meta.

That’s it for this week!

Stay experiMENTAL,

Sundar