How to Choose the Right Marketing Channels Using Data | #638

Most ecommerce teams are still arguing about channel split the same way they were five years ago. Meta versus Google, who gets the budget, who gets the credit. The argument usually gets won by whoever's most confident, not whoever's most right.

The deeper issue is the data underneath it. Every ad platform is built to over-claim, because the more credit it takes, the more budget you hand it. Add up what Google, Meta and your email tool each say they drove and you'll sail past two hundred percent of your actual revenue. They're all counting the same customer, so the debate gets run on numbers that don't add up.

The brands getting this right do three things differently.

In this playbook, based on a conversation with Paula Mitchell, Digital GM of Freedom, we cover three things ecommerce operators need to know about choosing the right marketing channels using data:

• Match the channel to the job the campaign needs done, not a fixed Meta and Google ratio
• Stop trusting the platforms' own numbers and judge the whole engine on one honest top-line metric like MER
• Buy independent measurement, but size it to your business instead of reaching for a full modelling platform too early

Connect with Paula Mitchell
Explore Freedom

Transcript

Copy

AI Search Offer From Studio Hawk

SPEAKER_06 0:00

Look, you probably know this already, but customers aren't just Googling products anymore. They're asking AI what to buy. And if AI can't understand your brand, it won't recommend it. That's why Studio Hawk, Australia's largest dedicated SEO and AI search agency, is offering at Descartes listeners a free SDO and AI search master plan. You'll get an in-depth audit of how your brand appears across the search and AI platforms, plus a custom roadmap to improve your visibility. And to help you get started straight away, you can also download their free e-commerce AI visibility toolkit. It's packed with prompts, checklists, wireframes, and a step-by-step digital PR framework. Head to studiohawk.com.au forward slash add to cart. That's studiohawk.com.au forward slash add to cart to download your exclusive toolkit and request a personalized master plan. We'll put the link in the show notes as well.

Why Channel Credit Breaks Teams

SPEAKER_06 1:01

Every e-commerce team that I work with has some version of the same argument running in the background and constantly. How much goes to Meta, how much goes to Google, and who gets the credit when the sales come through. Usually it gets settled by whoever argues the hardest or whoever has like an unwavering belief in a channel or just an old attribution model to fall back on. And the numbers that everyone is arguing with come straight off the ad platforms, usually, which is often the problem. If you add up what Meta, what Google, and your email tool each claim, it is more than likely that you'll land north of 200% of the revenue that's actually hit your bank account. It can't be right. Everyone is taking credit for the same sale. So the debate gets confusing and muddled really quickly, and the numbers often don't add up. Paula Mitchell has spent five years as digital GM at Freedom, running 50 stores and 70,000 products with a team of just 12. She has won just about every e-commerce award going, and she describes herself as the data girl who would happily hand all her budget to Google. But what changed her mind wasn't an opinion. It was independent data that gave her a surprising insight into how her individual channels were actually working together. So this week, I thought we'd unpack where your marketing money goes across channels, how to make decisions on what's actually moving the needle, and how to settle that argument with data rather than gut feel or strong arguments. We'll hear from Paula first and then from three operators who've each cracked a different part of this problem. So let's get into it. Let's hear from Paula first.

Mixed Media Modeling Changes Budget

SPEAKER_01 2:47

We've also recently implemented MMM, so mixed media modeling across the business from a brand marketing and a performance marketing. What does that mean, Paula? What's mixed media modeling? So we're able to, every month we refresh the data and we're putting in all of our channel data across our above-the-line channels, outdoor, radio, TV, as well as all our digital performance channels, all the social channels plus Google and affiliates. And then, you know, its algorithms and LLMs behind the scenes give you a view of, you know, when you increase or decrease certain channels, what impact it's going to have on revenue at certain periods of time. So if you want to be blown away and be drowned in data, but really cool data, it's been an eye-opener for the business to see what we think is working, that might not be working. You know, there's been chatter forever about TV being dead and you know, no one's watching TV anymore. But the screens model has changed now. You know, it's not just TV, it's all your B-bod and your S bod and everything that goes with it. It's just different kinds of platforms, I guess. People are consuming TV on than what they're used to. But, you know, the data that is, which is independent data, it's not done by nine or seven, it's an independent MBM platform. It definitely tells us that TV is not dead. And I'm the digital girl, right? So I'm going, give me all the money for Google. Don't worry about TV. And it's been a learning, right? We've got to find the balance. They all have a role to play, and it's where that part is in the funnel you want them to play, and how you dial them up and dial them down at certain moments. So that's been an awesome project between the digital and the marketing teams we've been working on together, just to help us settle the debate around budget split and uh how we just we can play nice in the sandpit now with data rather than opinions, which has been cool.

SPEAKER_00 4:33

What is your gut feeling then? Like if you had to like gung-ho go to your team and be like, if everyone was like, you can only pick one channel, what would you be putting all your eggs into? Like, which channel is it? Are you on the MetaDrug? Are you Google? Like, where are you?

SPEAKER_01 4:50

I think that's that's a suicidal question, I reckon. Anyone who has the balls to do that?

SPEAKER_05 4:56

Rosa, I've been doing this for six years and I've never had someone call my question suicidal. This is your third episode, and you've already got a suicidal question.

SPEAKER_00 5:04

This is a dramatic podcast.

SPEAKER_01 5:09

I would never have the confidence to do that unless I was 50% up on budget and everything was flying, and I had all the money under the sun and I could afford to not have any sales for a week, then I would, because I am the data girl and like to have facts rather than opinions, is I'd run one for a week and then another one for a week and another one for a week with all the same promotion and all the same to make sure it's a level playing field. What's your ratio?

SPEAKER_00 5:38

Like, do you feel like with freedom you've got to have like 60, 40 Google or Meta? Like, where do you think it's more important in digital for your budget to be? Like what channel? Because we talked a lot about attribution and how much insights you're getting from all this mixed media modeling. But what does that mean in terms of like budget ratio? Like, where are you focused? Yep.

SPEAKER_01 6:01

So you probably hate my answer, but I think it depends on the campaign. And I'll give you some examples. So coming into Christmas, we run entertaining. It's a Christmas campaign, it's entertaining, it's a table setting, it's inspirational. That belongs on Pinterest. And that belongs in a reel on whether it be Instagram or TikTok for you know a younger audience. So I think that's 100% where we would double down on those channels and we would take spotlight positions and we would up our budget, and we're gonna play in the upper funnel because we want to inspire people. But then when we're in flashstar mode or Black Friday mode, it's a majority in Google. When we're at the 11th hour and we're running a close and it closes in 12 hours, I'm double downing on you know, Google where people's intense signals are high, or even from a meta-platform perspective where we've seen that intense signal and we know they're already in the funnel, then we're doubling down there. So I think it depends always on your objectives and what you're trying to do.

Match Channels To Campaign Jobs

SPEAKER_06 7:00

I love here that Paula isn't naming a favorite channel. She's matching each channel to the job that the campaign needs to do. And she won't just trust a split that she can't independently check. So let's get into our three lessons around this topic. Number one, you've got to match the channel to the job, not a fixed ratio that applies across everything. Stop running every campaign through the same Meta and Google and TikTok and above the line and every other split. Stop just giving it one split across everything. Not every campaign is the same. Question that matters isn't your channel ratio. It's what is this specific campaign trying to do? Is it about doing an upper funnel job? Is it about introducing a brand new concept? Is it about changing an opinion? Is it trying to get them inspiration? Different channels will do different jobs. Pinterest, Reels, TikTok, Google search. They're all very different audiences in very different mindsets. Where teams often get caught is that they're holding each channel to the same outcomes for every campaign. But it's actually the campaigns that are different and the channels that need to adapt, not your campaigns fitting into the channel mix. Muscle Nation's Amelia Murphy is about as data-led as paid media specialists get. And he makes the same point about not judging every channel by its last click.

SPEAKER_02 8:23

Not everything has to be profitable last click, because we understand that if we're looking at it holistically, and it comes back to my point where I was talking about having a single source of truth, you're able to understand some platforms, for instance, like Google, inherently are going to have a really, really healthy last click because people are going to the platform. They're generally mid-funnel, so it's going to have a good last click. So that does a lot of the legwork when some other channels aren't necessarily performing as strong last click, but that's okay because you understand it drives traffic to mid-platform or mid-funnel platforms like Google. So it's all about looking at data in context, I think. TikTok traditionally isn't a platform that converts as strongly as others as meta, because inherently it's a platform designed to keep you on the platform for as long as possible. I know I'm guilty of it, I know it, I'm spending two hours a night on TikTok. From there, we're able to use that as not only a conversion platform, but a way to measure impact from other platforms.

SPEAKER_06 9:22

So before your next campaign, write down the job that you want that campaign to do. Then have a look at the channels available at your disposal to see which ones best fit that job and assign the right ratio for that campaign, not the other way around. And that brings us

Stop Trusting Platform Attribution

SPEAKER_06 9:40

to lesson two. Stop trusting the platform's own numbers. Every ad platform, or actually every marketing platform, whether that be email, ads, or other things, are actually built to overclaim. They all want to take the credit for what they do because the more credit you give it, the more budget you give it as well. So you've just got to be careful where you get your numbers from. The better approach is to have your own dashboard and be really clear around the metrics that you're tracking rather than relying on someone else's dashboard or metrics that they say are important. Here's a check that you can run today. Take your actual revenue from your store, then add up what Google, Meta, your email tool say they drove. You'll see that it will quickly sale past 100% because they're all counting the same customer with different attribution models and different attribution windows. That should highlight whether you've got a problem. Ian Calvert from Boom runs that exact test probably more times than anyone that I've spoken to. And his number is a really important metric that's worth remembering.

SPEAKER_04 10:40

Take all the sales that your back-end system that reports the money, that the boss is, then go in Google and see how many sales it thinks it's made. Then take Meta and see what it's claiming. Then take Clavio, if you're really going for it, like take just any channel that's reporting and trying to get your money, add it together, you're probably going to be about 250% off your sales. You think about what attribution is trying to measure. What made somebody do something? Five grand of like ads to get sort of like stuff going, unless you've got a massive sort of like audience, and five grand just all the other stuff. I'm known sort of like for e-commerce and all these different sort of like areas, but it's because I'm just genuinely interested in it. It's sort of like I like talking to people about it. And I'm fortunate that I can then go and sort of like do these things.

Right-Size Independent Measurement

SPEAKER_06 11:26

Which brings us to lesson number three buy independent measurement, but sized to your business and your stage of business. Once you accept that platforms can't referee their own game, you need a source of truth that sits outside of them. When Freedom put mixed media modeling in, it ended a long-running internal argument and told them with independent and agreed data that TV wasn't dead. Independent data earns its keep exactly when it contradicts you. But you might need to right-size it because not everyone can jump straight into a full media mix model. They actually take a lot of time and a lot of investment to get right. At the same time, not everyone should be running off an Excel spreadsheet that you're exporting from your platforms into one source. There might be something in the middle that pulls all that data together into a SaaS platform in the middle, which is kind of ready to go and built for the majority of the audience. There are so many different options when it comes to how to collate that data outside of the platforms. There are so many options when it comes to collating the information for an independent view outside of the ad networks platforms that you have to make a decision around what's right for the size of your business, how much you're willing to spend, and how capable your team are of using that data to make decisions in your business. Adam Sharon Zipsar talks about exactly that. As an agency owner, he sees this come up all the time. And his advice was not to jump straight into those big mixed media models because actually it might not be right. And it's probably not right for 90% of e-commerce businesses. So when it comes to setting up your independent data to measure the effectiveness of your media, you need to decide what your independent source is and what's right for your business right now. It doesn't mean it always has to be that source, but it has to be independent.

SPEAKER_03 13:21

Like if they're spending over $250,000 a month, there's a whole suite of like media mixed modeling tools and it's a massive statistical sort of framework. But then if you're under that spend, which 90% of the retailers are, look, we've been looking at incrementality testing. So changing one thing and understanding the impact off everything else based on your north size, your branded impressions, your traffic, and your revenue. And then looking at more aggregate numbers like an MER and pacing your MER over time. That's the modern way to do marketing right now. And that's where we're really understanding it.

SPEAKER_06 13:54

If there's one thing to take out of all of this, it's that the channel question and the measurement question, probably the same question. You can't decide where the money goes until you have a number that isn't being muddied by numbers coming from a platform with a bit of a vested interest. You need to match the channel to the campaign job that you've got on hand. You need to judge the whole engine on the metrics that are most important to you, not what someone else says are important. And you need to bring in or collate independent data to make those decisions in a consistent and considered

Join The Ad To Cart Community

SPEAKER_06 14:33

way. If you are trying to work through your own channel split right now and make decisions on where you should be spending your money in marketing and what channels are working for you, come in and join the Ad to Cart community. We have hundreds of e-commerce operators talking about this kind of stuff every day. We would love to have these conversations in there. Bring your problems, bring your ideas, bring your solutions in there. You can join for free at adducart.com.au. That is the playbook for this week. I'll see you next Friday.