Every week there’s a sensationalist headline about how the tenure of the CMO is going down.

I’m sure there are a variety of reasons but one broad reason has always been that Marketing has struggled to prove their impact in a way that’s credible for other functions. This has been especially true with teams like Finance and Product.

Having spent a decade in Marketing Data Science I’ve earned a few scars from having to continuously fight the good fight that Marketing is worth it so I thought I’d share a few ways Marketers can build credibility with data through some fun little conversational roleplay.

Throughout the article, you might be asking “but who am I gaining credibility from?”.

The answer is everybody:

Marketing is in the hot seat with everyone (which also is what makes it fun when you go boom look at these results).

Through a Marketer’s journey to create a strategy and campaign that delivers virality + business success are a series of check points that also act as points of failures for credibility. Each one of these is scrutinized but can be overcome with data that you can use to build up your credibility meter and cash it in when appropriate.

At the core of every good marketing campaign is a consumer insight. That insight can be quantitative or qualitative but it should be the foundation of any campaign or strategy. This sounds obvious but the reality is that doesn’t happen as often as we think. We know that because there are so many campaigns we see out in the wild and we wonder “What the f*ck were you thinking?”. You’ve also likely been in many of these conversations where someone proposes a cool new idea and it gains momentum but there’s no clarity on who the problem is solving for or what it’s solving.

Let’s look at a common conversation around insights that reduces credibility:

“Why are we running this campaign?”

“Because we think it’ll go viral”

“Okay, but what problem are we solving for the customer?”

“I don’t know”

The problem here isn’t that the marketer admits they don’t know. That’s okay. However, the expectation is that the Marketer does and so when expectation and reality don’t align you have dissonance. Especially within Consumer Tech, where the perception is that companies are data driven, Marketers must use data to find insights and becoming more technically savvy is a good way to increase credibility.

That doesn’t necessarily mean doing the analysis yourself but learning to better interpret data or even ask the right question is now table stakes. In addition, with the increasing adoption of AI, which reduces the need for an analytical counterpart, Marketers must learn to ask the right question and then surface them in a narrative way. Specifically, what Marketers need to unlock from an insight is what specific change in behavior leads to a change in growth:

Example of good insights

“When a user takes 2 trips, retention goes up 10%”

“When a user wait more than 5 minutes, their churn goes up 5%”

Now let’s revisit our previous conversation before and see how having insights builds credibility:

“Why are we running this campaign?”

“Because we think it’ll go viral”

“Okay, but what problem are we solving for the customer?”

“We saw that when a user does X their retention goes up Y% so we’re using this campaign to highlight this value prop in the attempt to get them to do X”

This is a much different conversation.

The pre campaign phase is often the most crucial one and can be a strong predictor of campaign success. When I say pre campaign, I mean a variety of activities:

Prioritization

"Why are we running this campaign?

“Because we want to focus on increasing awareness”

“Um, don’t we already have really high awareness. Right now we need to focus on retention”

That’s just an example (and maybe not the best one) but when cross functional teams are pulling in one way and you’re pulling in another that lack of alignment causes friction. So using data, Marketers can measure impact against the priority metric as a way to filter out campaigns that are not driving toward the broader business goal. This naturally builds credibility because when the campaign is aligned to the goals then the results will be too. Pretty straight forward but I call this out because I’ve seen a few times when the results are great but it wasn’t a priority and people question if the investment was worth it. By using data as a way to measure impact and ensure you’re laddering up towards priority you build credibility.

Let’s revisit the conversation that builds credibility:

"Why are we running this campaign?

“Because we want to focus on increasing retention”

“Awesome! How can I help? Should we jam on ways Marketing and Product should build together?”

Creative pretesting

"How did the campaign do?“

“It completely flopped!”

“Did we test it beforehand”

“Nope.”

“Why not?!”

Creative pretesting especially at larger companies and campaigns is a prerequisite and a huge risk mitigator. There are tools like Ipsos but they’re often expensive. So, what do you do if you’re a smaller company?

Some platforms have creative pretesting built in to the design. When you think Meta, you think about uploading 100 creatives and then letting the algorithm figuring itself out. That in a way is creative pretesting. I also recently spoke with a Head of Growth who uses TikTok as a way to pretest messages. She’ll just post on their organic page and see the reaction / engagement with it. She’ll test out multiple ideas and use engagement as a proxy signal for effectiveness.

Does it spark engagement? Conversation? Negative reaction? Good! The worst thing you can have is no reaction.

Let’s revisit the conversation that builds credibility:

"How did the campaign do?“

“It completely flopped!”

“Did we test it beforehand?”

“Yeah, it really did well in testing!”

“Has this happened before?”

“No, not that often. Usually pretesting is a good predictor”

“Okay, cool, weird…”

As Marketers we have to be comfortable with failure, but we have to mitigate the failures that are caused by us (internal) vs those that are out of our control (external). Creative quality is 100% within our control.

Measurement strategy

This topic is near and dear to my heart. It’s where the difference between having a good data partner vs not has the most impact. Measurement strategy is the #1 way marketers can build credibility before a campaign launches because it highlights 3 things:

This is the absolute worst conversion for your analytical counter parts that reduces credibility:

“Hey, we’ve got a campaign launch coming in 2 days”

“Okay, how are you planning on measuring this?”

“I was hoping you’d tell us”

“Um, well we’d need to hold out about 15 cities as control”

“Yeah, it’s too late for that… media buy is already locked in”

“Okay so yeah… we can’t really measure this “

The truth is that measurement requires sacrifice and planning and there’s an actual strategy behind it. Every data scientist wants to help prove that Marketing is working because that’s who is paying their paychecks. If you’re on a sinking ship of Marketing then you’re not really going to be there for long.

Here’s what a measurement strategy conversation that builds credibility looks like:

“Hey, we’ve got a campaign launch coming in a month”

“Okay, how are you planning on measuring this?”

“I was hoping you’d tell us”

“Okay cool. I propose that we pick these 15 cities (X,Y,Z) ”

“Ahh but those have new product changes so let’s swap out 3 other cities”

“Okay done. Here’s your list. Please keep these cities dark so we can do a proper treatment vs control”

These are conversations I’ve had over and over in my career and it’s not only about establishing credibility with other functions but also with your analytical counterparts. When data scientist know we’re being included in the conversation we tend to go above and beyond.

So, align on measurement strategy before hand. Include it in your briefing process. Make sure your campaign is airtight so that when you’re asked for results you can provide them.

Post campaign reporting is one giant set of traps because they’re full of ego and biases. Here’s what I mean:

“Okay, how’d the campaign do?”

“We saw an uplift in 4% in signups”

“Wasn’t the goal supposed to be new orders?”

“Yeah it was.”

“How’d that do?”

“Yeah there was nothing there…”

“ummm….”

“But look… signups!”

If you have a measurement plan set up then stick to the measurement plan and share what the results show. Don’t try to do measurement theatre. It’s short term good for long term bad and Marketing is one of the worst offenders of this.

There are also a few other ways to build credibility as a marketer. Here are some key things you can do to gain credibility with data:

A bonus way to build credibility is to then take the learnings from the one campaign and turn it into multiple experiments. I saw this a lot with Lifecycle campaigns where you could segment the audience after a campaign and see that different segments performed differently. So, you then launch a new experiment redesigning just for a smaller segment and potentially see uplift there. This also builds credibility within the marketing org because you’re showing continuous evolution. There are so many moments to build credibility!

Let’s revisit the post analysis conversation that builds credibility:

Okay, how’d the campaign do?”

“We saw no stat sig uplift in new orders but did see a stat sig uplift in signups”

“Okay, cool, thanks for sharing! What are you planning to do?”

“Well, we want to dig into why signups didn’t convert to new orders so we’re prioritizing digging into that and hoping to relaunch a test within a few weeks”

Optimizing is such a generic term but what I’m referring to here happens across 2 pillars: channel investment allocation and cross channel budget optimization. Let’s dig into both of those:

Channel Investment Allocation

Channel investment allocation refers to how much you should be investing within a channel. It’s a common question and one that can most certainly be answered by data.

For example, on Meta let’s say you’re currently spending $100 and it gets you 10 Signups. What would happen if you double it? You spend $200 and it gets you 17 Signups. What if you halved the budget? Tripled it?

What happens is you begin plotting a cost curve and you’ll always find a diminishing return curve where the additional amount you spend doesn’t have as good of a return. Using data (and data science) you’re now able to find the most efficient point on the curve of where you should spend. Think about how much credibility you can buy!

Cross Channel Budget Optimization

This takes us from the more campaign level focus to the cross Marketing discipline on how to build credibility with data. We enter a meeting room and hear the following:

“Why are we spending money on TikTok?”

“Because we’re seeing a decent return”

“Okay well about if we took some of that and went to Meta?”

“Well we’ve tried that but don’t really see impact”

“But even like a little bit”

“Hmmm maybe”

So once you’ve got data to prove clarity on the individual campaigns and channels, you can use data to ensure you’re spending $ the most efficiently. This is often done through tools like incrementality testing and MMM but without data (or data science) you’re likely giving a large % of your budget to Meta / Google.

The quickest way to your CFOs heart is through an incrementality test.

Here’s that same conversation but it builds credibility:

“Why are we spending money on TikTok?”

“Because we’re seeing a decent uplift of Y% incremental at a $Z ROAS”

“Okay well about if we took….”

“… some of that and went to Meta? Yeah, we thought about that and found that we were overspending on Meta so we’ve allocated $Z from Meta towards TikTok.”

None of these efforts are easy, but they 100% build credibility and in the long run credibility + trust are the paths of least resistance.

The best way to build credibility in a modern Marketing world is through the use of data. You can use data at multiple points of the journey:

Each step builds more credibility and collectively makes a Marketing team, function, or org irreplaceable.

Note: Much of this is higher level but over the next few articles I’ll be going deeper on some of these topics including how to scale channels (like meta) and how to better experiment along with case studies from my time at Uber about how we tackled these challenges.

Stay experiMENTAL,

Sundar

Share with your colleagues or on your LinkedIn . It helps the newsletter tremendously and is much appreciated!

Here it is: Why analyzing Marketing feels impossible