On Brand No. 15, featuring Chase Mohseni of Pencil
Table of contents
Building Brands

On Brand No. 15: What makes a good ad? High production value? Maybe not

September 08, 2022 · 9 min read

This is an edited version of an email newsletter sent on August 24, 2022. We send out new issues every other Thursday, featuring deep-dive essays and interviews with industry leaders. Sign up below.

As of September 8, 2022, On Brand is on hiatus. We have closed new subscriptions for the time being. Read the backlog right here!

👋 This is the last issue of On Brand for now. Why? It’s simply a matter of resources. We’re not a large marketing team, and other projects are demanding my attention. I’ve thoroughly enjoyed making this newsletter — meeting new people every other week, reading the replies, creating silly little graphics. Thank you, truly, for reading along.

— Francis, Content at Air

This week’s plot:

  • What doesn’t make a good ad?

  • Spicy takes on the world of digital advertising

  • What’s with all the fuss about AI tools?

Chase Mohseni: A wealth of digital advertising knowledge (and hot takes)

For the past couple of years, Chase Mohseni has been working as Pencil’s Head of Marketing and Growth. If you haven’t heard of Pencil, here’s how Chase describes it: Pencil helps you scale your business through creative testing by automating the entire creative production process.

Safe to say that Chase has a wealth of knowledge to share regarding digital advertising. He’s a triple threat: smart, humble, and generous. You have no idea how much of this interview ended up on the cutting room floor — he had so much to say. I guess if you like what you read here and want more of what Chase has to say, you’ll just have to subscribe to his newsletter.

Chase connects the dots at Pencil
Chase connects the dots at Pencil

On Brand: You used to be in film — directing, writing, but mainly producing. What did you learn there that informs your marketing work today?

Chase Mohseni: Every business has a story, regardless of if they’re selling $10M enterprise deals or $2 cans of soda. A collagen company might sell 10,000 products per month, and a movie studio might sell 8 movies per year, but both businesses need to tell a story. The only real difference is the volume of people they’re talking to. A movie studio is high-volume, low-touch; a collagen company is low-volume, high-touch.

It just comes down to talking to and connecting with people. The people at the business — what story are they trying to tell? The people you’re selling to — what’s their story, how does what you’re selling fit into it?

Lastly, in filmmaking, you’re always editing. Tinkering, tinkering, tinkering until you eventually just have to put it out. It’s similar in marketing: you want to iterate as much as possible, but eventually you need to let the world tell you if what you’ve been working on is good or if it sucks.

OB: What’s a common misconception about what makes a “good” ad?

CM: People think everything needs to be highly produced. I love a pretty ad, but I find that pretty ads are made more for the marketing team than for the customer. To solve this, I’d push people to actually talk to your customers. Product teams always push to have interviews with people — marketing and brand teams need to do the same thing. Not a survey! Actually go talk to people, that’s how you avoid falling for your own biases.

Pretty ads are cool, but ads that convert are better. Everyone uses UGC, which is great, but you don’t necessarily need to have someone talking while waving your product around. Make some lo-fi videos just showing your product, even use static images within your videos. It doesn’t need to be highly produced.

Chase is an avid tweeter — here’s one of his best.
Chase is an avid tweeter — here’s one of his best.

To be clear, this is vertical specific. In fashion, it’s a little harder. It’s also AOV [average order value] specific. If you’re selling something above $250, there is definitely an expectation of slickness. Do the research to find out if your customers expect higher production value. But if you’re selling something for $50, you probably don’t need to worry about it too much.

Don’t get caught up trying to imitate massive, established brands, either. They already have market trust and don’t necessarily need their ads to convert. Look to them aspirationally, but be honest about the reality of your stage of growth, and what you need to accomplish.

OB: What’s the hardest part of digital advertising work? Where do the biggest breakdowns happen, even before an ad is published?

CM: It depends on team size. Let’s say you’re on a team of 10-15 people, total, and you need approvals. The breakdown will happen with communication. Agencies over-index on this in a good way, where they have incredibly thorough, frequent communication with the teams they’re working with. But then an executive might see an ad the agency has produced for their company and say, “I don’t like that.”

The thing is, you need to run something. The executive team might not understand what you’re trying to do and why — which is because you haven’t briefed them on your goals and methodology. Make sure everyone has the necessary context.

The other issue is you need a testing strategy. Look, running a business is hard. Marketers wear so many hats. It’s more work. But you need a regular cadence for ad testing. I’m not just saying this because I work at Pencil, it’s just a statistical necessity: you have to run a certain amount of ads to find winners. Those winners bring you incremental profit. They bring you new customers.

How Pencil thinks about ad testing
How Pencil thinks about ad testing

So it’s two things. You need thorough communication and a clear testing strategy — those are the two most important parts of digital advertising work to get right.

OB: Give me your hottest takes on digital advertising.

CM: One, people over-index on the ads they like versus what the customer likes — what will actually convert. Don’t do that.

Two, paid and brand people often work against each other, think they’re enemies. But without a brand, the growth team won’t be able to build a sustainable motion. Without the growth team, the brand team can’t get the money they need to operate. Brand smooths out the valleys that inevitably happen in a growth cycle. Stop fighting, you need each other!

Three, people don’t actually know what growth marketing is, they just say they’re in growth. Growth is predicated on repeatability, making sure you can grow the LTV [lifetime value] of a customer, and turn them into a marketer for your business, bringing in more customers “Growth” is a buzzword people like to throw around without knowing what they’re talking about — I was one of those people for a long time!

Fourth, people are way too addicted to quantitative data. You need to synthesize qualitative and quantitative. Qualitative feedback from customers allows you to understand the narrative through-lines in your data. I call it narrative data — you only get this by interacting with human beings. Every datapoint on your dashboard is a person with a unique experience. Learn their story, learn their context, and it will give deeper nuance to how you market.

OB: People worry that AI tools — like Pencil, but also Dall-E, Jasper, and the rest — will make skilled workers redundant. How can people take advantage of an AI tool like Pencil to empower themselves, rather than becoming redundant?

CM: It’s nuanced. Any time something new comes out, the Twitter of it all produces a bunch of un-nuanced opinions. Look: Pencil literally cannot work if you don’t put assets into it. Who creates those assets? We don’t do that at Pencil, nor do we have ambitions to do so.

You still need the brilliant people working upstream to make the assets. But ask those talented designers: do you like making all the ad variants? All the different placements and tweaks? I’ve never talked to anyone who says yes. Nobody I have ever met says they enjoy making yet another Facebook ad. It’s just part of their job.

Pencil simply compresses that part of your job so you can work on the stuff you like to do, like a landing page, or creating a cool brand campaign.

AI, at the core, is made, analyzed, and retooled by human beings. All it is is a faster brain for synthesizing data and information; it cannot create things without inputs.

An example of what Dall-E 2, the image-generation AI, can do with an artist-created input.
An example of what Dall-E 2, the image-generation AI, can do with an artist-created input.

Take Jasper — you give it inputs, it spits something out, and you aren’t going to use exactly what it writes. You will retool it, fix it, work with it.

By the way, everyone who uses Facebook ads? That is an AI, and you don’t question it because you don’t have a choice. It’s a tool that makes your job easier. It used to be, you throw some ads on TV, on billboards, and you wouldn’t be able to get nuanced insights about their performance. Facebook and similar advertising algorithms changed the game for marketers — that’s what AI tools like Pencil can do, too.

At the end of the day, AI is just meant to compress tedium and make things smoother for you so you can do the work that actually moves the needle. Is spending five hours on ad variations really the best possible use of your time? People use tools to make their lives easier, that’s all it is.

OB: Your team created a brand, The Pencil Company, to dogfood your actual product, Pencil. Tell me about that process.

CM: Our CEO Will Hanschell and I came up with this on a late-night phone call. We were thinking — what can we do that can get us closer to our customer and allow us to have more meaningful conversations on sales and marketing calls?

So we built a little DTC brand. It was cheeky — as Pencil, we should sell pencils. We had a website already, pencil.co. We started a Shopify store, used Soona and Billo to get some assets, sourced some product from Alibaba and Etsy, and created a store with real products.

We ran our own Pencil playbook, created some ads. The problem was that the margins on pencils are really thin, so running ads was a little dicey — we spent a few grand on ads, and realized it wasn’t worth it. We’ll probably try again with sweatshirts, where at least we can make $20 a pop when we sell them and reinvest it.

What did we learn from this? Sometimes we’re in a vacuum and don’t realize how hard it is to make a sale. It pushed us to look at the bigger picture and where ads sit within it. We realized questions we should be asking our clients, concepts we can put in our ads to connect with prospects better. We also realized there were better ways of qualifying our customers. Before, we thought we could just help anybody — like, if you want to go to one of our higher pricing tiers, these are the specific things your business needs to support that and get the most out of it.

It was a great experience, but we got a little caught up in the joke of selling pencils because we’re called Pencil.

Key takeaways ✏️

  • ABT — Always Be Testing. Digital advertising is all about iteration. Have an incremental testing strategy and stick to it.

  • Don’t get caught up on the production value of your ads. Run customer research, create ads based on that, then run them, learn, iterate, repeat. Published is better than perfect.

  • Don’t over-bias towards quantitative data. Qualitative research is necessary to understand the human reality of all those datapoints on your dashboard and build a narrative.

  • Don’t fear new AI tools. They’re meant to compress your work and automate manual tasks — they don’t replace talented people.

In the know

Must-reads, hot takes, and rising trends:


Marketing + Creative jobs at our favorite brand-forward companies:

Air is also hiring for a variety of roles! Click here to apply.

Related posts


Why we sent 30 elderly people to protest an Adweek Conference

March 05, 2024

At Air, we’re always trying to find new ways to engage our audience and meet them where they are. But only rarely do we try and take that last part literally. When we read that Adweek would be hosting its Commerceweek event in New York City on February 28th, we decided to try something a bit… different.

Unblock creativity.

Everything you hate, we automate

Need to talk to a human first? Request a demo →

Air workspace