How To Approach Experimenting With Growth Marketing


Table Of Contents


Why Is Growth Marketing Important? 

Growth marketing focuses mainly on building customer relationships and fostering loyalty, and Nima Gardideh is an expert on this. As the co-founder and CTO at Pearmill, he works to enable better digital performance through advertising and creative marketing. 

Utilizing tech to create a strong digital presence is vital in today's marketing landscape. In this blog, Hollywood Branded shares how to best approach experimenting with different aspects of growth marketing to find what works best for you. 

EP317 How To Approach Experimenting With Growth Marketing

A Little More About Nima 

Nima is President and Chief Technical Officer Pearmill, a tech-enabled growth marketing studio. He and his team work with brands to enable better digital performance, while also focusing on advertising and creative stories to artfully market a brand. Having previously been in the position as the Head of Product at Taplytics (a YCombinator company), the Head of Mobile at Frank and Oak, and at a venture-funded consumer company, Nima is well versed in the start-up world. Today, he works with leading technology brands to help them acquire customers through art and technology by bringing growth experimentation process targeting, creative production, and data engineering to major digital platforms. As a digital nomad, who lives life as an artist, entrepreneur, and engineer, one of Nima’s core beliefs is that humans are influenced by narratives and that stories have the power to evoke emotion or thought, regardless if based on science, culture, or fiction.

New call-to-action

Interview Transcript Highlights

Question: How did you get started doing what you're doing?

Answer: My background has been really in technology, I started as an engineer and moved into product management later on in my career, and what happened was that we started a different company, it was a consumer company to help people buy products better like to think of it as consumer reports of the future, and we learned very quickly that that business model work, but what we learned as part of that was how to run marketing very well, so we started consulting with some folks that we had gone through y combinator with and helping them acquire customers and find your audience and over time, that became the business that we're running now.

Which is sort of a combination of I would say, four different disciplines, there is Growth Marketing so media buying where are we putting the budget and allocating the budget, Creative Production, what are we showing these users and how are they seeing in what context and what channel, etc. Conversion Rate Optimization, so what is it like after people click on an ad what is the experience for the user, and how are they going through the flow and becoming a customer. The last part is Data Engineering, as you mentioned, just kind of a combination of the glue between all these things. What is the data, how are we looking at the success of the work that we're doing, how are we understanding, where should the money go where should the next dollar go, and how can we make sure that we are spending the money in the wisest way possible right and so there's technology on that front and sort of in every single part of this, there were some app or process that we've created automation around that and in our company, and so our backgrounds, even though they're in technology have merged with this world of marketing in a different way, I think, then maybe the average marketer ends up being, and I think because of it, we have some advantages and, of course, some disadvantages that over time we've been able to hire for to help us get better at those spots, but that's how we've ended up here.

Question: So when people think about growth marketing performance marketing, what do they think will happen versus what actually happens with that type of marketing practice?

Answer: I think there was there's basically a school of thought that has existed for the last maybe let's say 12 to 15 years that's changing dramatically right now, as we speak, right so.

Let's say the growth marketing performance marketing world it really started becoming a bigger portion of people spent after mobile ads became a thing, where you could have perfect attributes right, so I could spend X dollars on Facebook and I would see X times five out of my APP data within a certain period of time.

And so, then I can make very clear calculations about where my money is going and how much is going to return, so it was like a dollar in, dollar out reality when performance marketing is done maybe past marketing brand marketing and previous forms of marketing didn't have as tight of a grasp on the obvious obviously care everyone in marketing cares about the value of the work that we're putting in, but it was harder to calculate it was harder to extrapolate what is the exact value that we're getting out of this app, right?
That's changing now, because Apple has changed the rules of the game where the data is not coming in, as clearly anymore, and you now have predictive models for calculating value, so we are moving to a middle ground where we still get a sense of the value, but we don't get the exact value anymore.

At least we don't get it up on a per person basis it used to be, that I if you clicked on an ad I knew exactly what ad you clicked on.

And then, six months later, if you spent $1,000 with our brand and I could extrapolate exactly how much it cost me to get you to click on something and buy our products that are no longer the case, I would say, I can tell that you were one of 10 or 30 customers that clicked on an ad that may have looked like this on that date that's kind of it's more of a predictive we get a sense of that you're part of a group of people that had performed a certain action and, over time, has provided this sort of value for us, which I think is better for consumers, I think that level of accuracy was almost too much understanding of your user base.

I think I'm generally pro-privacy, and so I think it's wonderful that the industry is moving towards an area where it's not as predictive in terms of how accurate data is, but we still have information, so we can be good stewards of the capital, I think the biggest hit unfortunately for these ios 14 changes, has had has been on the early stage companies, the ones that are not spending a lot they don't have a lot of aggregate data to look at to make better decisions with if you had scale, the problems i'm talking about or or less because knowing that you are one of 10 customers that clicked on this ad is actually enough for me to make good decisions.

But if in total I'm getting 10 customers a week, I'm not getting really any data to be able to make good decisions so the changes have hurt small businesses, but ultimately are better for consumers, which is the trade-off that I think these privacy policies of the next maybe five years are going to have to make is where can we provide the value for SMVs and working we provide more value and more privacy to consumers.
Extreme no privacy was giving a lot of value to SMVs, but now we're kind of balancing it out.

Question: When you're doing this and you're trying to live in this land that's friendly to customers, but actually works so well for your clients, what are some of the issues that come up what are some of the mistakes that happened in the approaches?

Answer: If we go back to the process conversation…I think if you're going around reading about growth marketing and how you can do the best job being a growth marketer… Everyone will talk about how you have to be experimenting with things, and you have to run experiments and come up with ideas and try out different audiences, and try different creative. And you know if you're a slightly larger scale try out different data optimization points like what events are you trying to optimize towards and things like that, right? Yeah and I think that sounds easy. All I have to do is try a bunch of things, the problem with it, is that it's very hard to get real learnings out of these experiments. So let's say, I want to try out a different experiment. So, Facebook has this feature called the look-alike right, so you can grab an audience and create a lookalike audience and you want to see if that's a performance audience. How are you going to actually run that test? Are you going to create a different campaign? Are you can create a different ad set.

There are essentially three or four different ways of running this test, and then you have to isolate the data in a way that you can actually trust that ‘Oh, when I ran this test, improve the data by 20%’ and how much money you put part towards it is actually a variable.

So, if you put enough money behind that maybe it will work, but if you're risking too much by putting that so there are all these variables, you have to think about.

So, what I think is the hardest part is this process, how do you actually run this process, and how to instill it within your team to stick to the process. 
So for us, the way we've solved this, for now, at least is that we have a very clear way we run experiments or every part of this, we have at least one or two approaches, we take. And we stick to that approach right so for audiences we test them in a certain way we put them let's say in a specific outset we allocate this percentage of the budget to it and that's a variable of how much risk we're willing to take on the test, and the same thing with creative, how are you actually allocating the budget to creatives. What is the structure within the Ad account for how to run these creative tests, who's deciding which creative ones and which ones didn't win and how can you extrapolate the learnings, right?

For us we've now gotten to a point where we can tell you, for instance, one of our clients is named Xander, there are at the distributed hotel, we now know that if you have issues of their apartments in such a way that there's a kitchen countertop into the image, there are massive windows and there were big high ceilings and the image, people are two or three x more likely to convert. We learned this over time and then all the learnings all those learnings are in a document where when the creative directors want to attempt, another test they know all the things we know about their customers and how can we use that to run new tests and get more insights and learnings from the account, so these things I think are hard to build within the organization and harder when you're actually in-house. Our whole job is to master this process right and we still have problems doing it yet because it's hard to get people to follow these things. Imagine being in-house, and you don't even have the resources that we have to care about these things it just is a hard thing to do, so I think if you can nail the process we have not ever failed at scaling up an account that has a good product. If we've stuck to our process, and so that's the biggest learning is, can you actually stick to one of these processes that people talk about and create rules and understanding of each ad network has its nuances, how do you test things within each app network like Facebook vs. Google vs. Tik Tok vs. Snapchat. And have real hard conversations about I know we ran this test, but it didn't follow our process, we cannot trust this data, we need to run it again. And that happens to us still and those are in a way, winning moments for us because it shows that our team cares about understanding what actually happened with their tests, even though that someone maybe made the wrong decision at that time when they ran the original version of the tests and ended up in a situation where we have to run the test again, right? Or sometimes the client pushes for too many variables and we end up with no learning there are too many variables introduced in the test.

Question: Is there an ideal number of variables when you're doing this? Are you trying to limit your team, so that there's like more of a focus versus the almost overload of ‘Hey, it's not a B testing, now it's A to Z testing?’

Answer: Yeah I think it's a good question, so I would say um it's very dependent on the scale right, so if we're at a massive scale, I would like to variable to be one.

If we're at earlier stages it's impossible to do one variable at a time, because you just do not have enough data, so what you want to be doing is grouping the variables in a way, that are related as much as possible, and there is low variance, right?

so if you're running a creative test let's say and you're running five different tests, the five should be as wildly different as they can be as possible. So, then you have now directional information about where to go next right so let's say you try animation vs. illustration vs. cinematography of this approach vs. static vs. video, and those are the five tests that you're running… Two of them do well there's almost no overlap that's actually fantastic write two of them did well now you have general areas to explore.

And then you cut it down a little bit more right so with animation you try different forms of animation let's say or different formats and different ways of writing the copy and things like that, so you just kind of slowly build up on top of it. But if you're at scaling up by skill, I mean you're spending hundreds of thousands of dollars a month.

Our clients don't on average you're spending, about half a million dollars a month, so we are able to test and much more focused ways.

Where we'll test one new audience and it’ll only do for a week and we get the answer within five to six days, because we have the scale and we can allocate the budget in such a way that we get the answer very quickly um yeah, so I would say, like yeah The short answer it depends basically.

Check Out The Podcast!

Nima has so much great information from his experience in growth marketing. Check out the podcast below to learn more about how to best utilize growth marketing. 

Every week we have a marketing professional on our show to share their tips, tricks and lessons learned from their professional experience. Check out some of our other podcast blogs from earlier this year: 

Every week we release a new podcast featuring guests with so much knowledge about marketing, you don't want to miss one!  How can you make sure you don't miss an episode? Click below to subscribe!

New Call-to-action