How to Become Antifragile and Survive Volatility

As 2020 continues to throw curveballs left and right, the only thing we know for sure is that we have no idea what is coming next. That’s a tough pill to swallow, especially if you are a business owner trying to plan for the next quarter, or even the next week. Consumer behavior continues to shift in varying directions, and every industry in the world seems to be in a constant state of flux. 

With so much volatility, what is an entrepreneur to do? Taylor Holiday has some ideas. Taylor is the Managing Partner of Common Thread Collective, an agency that helps eCommerce companies grow from zero to millions. Recently, the companies he works with have been forced to change the way they operate.

On this episode of Up Next in Commerce, Taylor takes us through what it means to build an antifragile business, and how that mentality can lead to a thriving business despite what the market or current environment has in store. Because, as Taylor says, there’s no point in trying to predict what the future holds, and instead, founders should be creating many different models so you can prevail even during volatile times. So what does that mean for your Q4 strategy? How should you be preparing for Black Friday and Cyber Monday? And what data should you really be looking at when developing a Facebook ad strategy? Find out all of that and more in this interview. 

 Main Takeaways:

• Never-Ending Qs for Q4: 2020 has been the year of uncertainty, and Q4 will be no different. Traditional planning for end of year events like Black Friday and Cyber Monday have to be approached with a new mindset, one that can adapt and pivot quickly. Companies need to put systems and plans in place so that they are prepared to take on any and all scenarios that might arise.

• Building Something That Works in Spite of You: Modeling and forecasting are tools that every business uses to help guide strategy, but neither are ever 100% accurate. Because humans are wrong more often than they are right, it is critical to set up systems that can survive not only when you’re right, but also when you’re wrong. That is the fundamental practice of being antifragile. 

• How The Past Predicts The Future: Drawing insights from historical ad campaigns is a double-edged sword.  When it comes to analyzing data, you can’t look too far back or too far forward. Yesterday’s ad data can help inform your decision for what to do tomorrow, but it can’t help you make a month or year-long ad strategy. What can be beneficial, however, is poking through the creative assets of campaigns from companies decades ago, pre-internet. Those are timeless sources of inspiration that can help you stand out from the uniform ad campaigns of today. 

 

For an in-depth look at this episode, check out the full transcript below. Quotes have been edited for clarity and length.

---

Up Next in Commerce is brought to you by Salesforce Commerce Cloud. Respond quickly to changing customer needs with flexible Ecommerce connected to marketing, sales, and service. Deliver intelligent commerce experiences your customers can trust, across every channel. Together, we’re ready for what’s next in commerce. Learn more at salesforce.com/commerce

---

Transcript:

Stephanie:

Welcome to another episode of Up Next in Commerce. This is your host, Stephanie Postles, co-founder of mission.org. Today, we have Taylor Holiday join the show, the Managing Partner of Common Thread Collective. Taylor, welcome.

Taylor:

Thank you so much for having me. I am excited to be here.

Stephanie:

Yeah, I'm really excited to have you here as well. I have followed your Twitter threads and I think we're going to have a lot to talk about today.

Taylor:

Okay. That's good. Now I feel accountable to everything I've ever said. So here we go. Let's see what I can dig up.

Stephanie:

Yep. I've looked at everything back to 2008, so we're going to cover all of it.

Stephanie:

I wanted to start with your background a little bit. I saw that you were in the world of sports.

Taylor:

Yeah, I did many [crosstalk 00:00:50].

Stephanie:

And I wanted to hear how you evolved.

Taylor:

Another lifetime ago, I was a professional athlete. I played baseball in the Yankees organization for a couple of years, and that was my life for the first 25 years of my life, was committed to that pursuit. Then one day I got a call and they told me that they were no longer interested in my services and I had to figure out what on earth I was going to do from there. That sort of set me off onto the second phase of life. I'd like to think I'm breaking my life in this sort of 25 year chunks. I'm about halfway through the second quarter.

Stephanie:

That's awesome. What did you decide to do after that?

Taylor:

Well, I didn't really decide much. I was finishing up, so I got drafted when I was a junior in college, so I had some school to finish. I was sort of in the off season. I would go back to school and take them at a semester at a time. When I got released, I started trying to figure out, okay, well, what was I going to do? And I was a political science major with a minor in psychology, and I loved to argue, and so I figured, well, I'll try and be a lawyer. That was sort of what I was prepping to do. I was prepping to take the LSAT and head off to be a lawyer. Then I had a good friend that had been a childhood friend, and is still now my business partner named Josh, who was starting a company. In between class, he would let me come to their office, which was him and his brother in a garage, and print the orders off the website and take them to the post office, and that was my job part-time.

Stephanie:

Sweet.

Taylor:

One day, day one, there was one order, and then within a year and a half we had done 60 million in revenue and that became my business school, and how I got into entrepreneurship.

Stephanie:

Wow. That is crazy. That's really good growth, and I'm sure you learned a ton while working there.

Taylor:

It was wild. But it sort of met everything in me that being an athlete did. There was a team of people that I love, working towards a common goal, every day you showed up and had something to do intentionally to be better. I was single, I was young, I had nothing else to do. We just lived at the office, and it was everything. As the business was sort of growing, we got asked this question of like, "Well, this is a real company. What's your job going to be?" And they went, "Well, you're the young person. Why don't we figure out Ecommerce, social media, and you know some famous people, so how about influencer marketing?" And I was like, "Okay." Then I started Googling, how do you set up a Facebook page. And just had no idea what I was doing, but learned, got to play in a sandbox, where suddenly I developed a skill that mattered in the world. I got really lucky in that sense that they entrusted me with that responsibility.

Stephanie:

That's great. What kind of famous people do you know? Now I'm intrigued?

Taylor:

Oh, so many famous people. No, not really. I had played professional sports, so I had a lot of relationships with athletes and agents and people like that. Our product was built for that community, and so it was just literally Facebook messaging friends and being like, "Hey, can I send you this product? Would you wear it?" That snowballed really quickly. We ended up building an incredible athlete team with ... at one point, we had all four MVPs of the major sports. We were brokering deals with Kobe Bryant and China, Shaquille O'Neal was a business partner. It was wild. We got involved in so many things in that first business that we had no business doing as 26, 27 year old kids, and made every mistake you could possibly make, but just learned so much that has sort of been the foundation for what we get the chance to do today.

Stephanie:

That's great. Yeah, that's a really good story. So fast forward to today. Tell me a bit about Common Thread Collective. What is it and what is your role there? What's your day to day look like?

Taylor:

Yeah. Common Thread Collective is an Ecommerce growth agency. We help consumer product Ecom businesses grow from zero to $30 million. That's sort of the range that we focus on, and we do that through a combination of paid acquisition services, email, SMS retention and landing page development, and then creative for that whole customer journey. So, we really see our role as the guide for our clients along that growth trajectory that we've lived ourselves and are currently living alongside them with the brands that we own and operate. So, we sort of approach growth from an operator's mindset, which we think really sets us apart from a lot of marketing agencies. My job is to be the CEO of that organization. It is certainly a very different job than when I started where I was doing the work. I spend much more of my time now thinking about organizational structure and culture and hiring than I do about marketing. That has just sort of been my own personal evolution, which I'm learning to love. But yeah, that's what it does.

Stephanie:

That's cool. So how do you go about picking who you want to partner with, which companies you want to?

Taylor:

In terms of the clients?

Stephanie:

Yeah, clients.

Taylor:

Yeah, so our mission for Common Thread Collective and really for our whole ecosystem, and I think Andrew talked a little bit about this on your podcast, is to help entrepreneurs achieve their dreams. That is our heartbeat. It's what drives us because we ourselves have experienced the transformation that comes from being a successful entrepreneur, what it offers you in life. So, we love to partner specifically with founders and entrepreneurs who are in that range of business. Usually, when clients come to us, they're somewhere between two to 10 million in revenue, and we're helping them sort of reach that next phase of growth along the way. That connection to the person who is passionately committed to the product or idea that they have is what motivates us. It's what keeps us engaged, because the reality is, when you work at an agency, you're not going to love every product that you're working on.

Taylor:

You're not going to care as deeply as the founder about hair care and sports wear and fitness equipment and beauty products and vitamins. There's just no way, but what we learn is that, what really works is when we care about the people. That's when our people internally are the most inspired, that's when they wake up in the middle of the night and think about ideas for the product, is that when they fall in love with the humans on the other side of it.

Stephanie:

Yeah, that's great. I want to jump right into something I've been following. You were discussing a little bit about how brands should be approaching holiday season, Black Friday, Cyber Monday, and how they should be thinking about their marketing and advertising efforts. I think it was on Twitter, maybe as an email thread, but I was hoping we could dive right into that, because we hadn't actually talked about that on the podcast so far, and I think it's a perfect time to kind of discuss how you think brands should be prepping for Black Friday and Cyber Monday and how it's different than in the past.

Taylor:

Yeah, absolutely. This is a crazy time, right? It's never been a more volatile moment in the history of Ecommerce, which is not necessarily the longest history in the world. I would put it right up there with every business season in our country's history, certainly in terms of the volatility of the moment. When you think about trying to forecast into an environment that is this volatile, there's huge error bars on any prediction that you're going to make as a business owner. If you think about some of the things that we're looking at, as we think about Q4, is retail going to be open? Are people going to be able to shop in stores? We have no idea? Is the USPS going to be able to handle the influx of demand on the infrastructure? We have no idea.

Taylor:

What is the social position of our country going to be after this election? We have no idea. As you think about that, what you have to sort of agree or accept is the idea that whatever you think is going to happen is likely going to be wrong. What that means is that, unlike years before, where we were in a more stable environment, you need to have plans that account for different possibilities. As you think about something like your Black Friday promotion, which in years past, maybe a very simple exercise of just going well, we're going to try and bundle some products and do a discount, you need to consider the possibility, well, what happens, what would our discount be if USPS doubled their rates? Would we still be able to offer and afford free shipping?

Taylor:

I would start to have multiple plans. The same is true for the tone of your messaging. If we come out of an incredibly hostile election on November 4th, and three weeks later we're in holiday season, and the country's significant unrest, what is the right message for your brand to have to sell products into that environment? Rather than trying to guess which one is going to be accurate, I would begin to have a multitude of plans for this moment in ways that we've never really had to consider that level of volatility before. That's one of the big things we're talking with our clients about, is this idea of how do you deal with the idea that you are most likely going to be wrong about whatever you think is going to happen in the future?

Stephanie:

Yeah, that's really good. I like the idea of making scenarios so you don't have to predict the future. How are you thinking about advertising? I think you were mentioning that you can actually prep in a way that you know it's going to be expensive to advertise during those times, and so how brands can actually start prepping early so maybe they're not being met with these really high costs.

Taylor:

Yeah. I would just contend that I don't actually know that it's going to be expensive. I think that's a theory that people have, is this idea that big retailers are going to be allocating a bunch of money into the platform and CPMs are going to be through the roof. But we have seen really dramatic things happen where like, last time when the pandemic got peaked in April, all of a sudden cell phone usage goes through the roof, the inventory allotment for ads goes really high, CPMs plummet. The idea that we know for sure, this is, again, sort of that idea of the contingency planning, I think is really hard. What that means, and I think what you're driving me towards is this idea that, how do I build revenue in a more predictable fashion when the ad environment could be incredibly volatile?

Taylor:

What I would say is that, when I think about forecasting, we described Ecommerce forecasting like a layer cake. The base layer, the foundation with the least variability is your existing customer set. You know that when you acquire a customer, they're going to produce future revenue for you as well as present revenue, and they do that really predictably. It's not subject to CPMs, it's not subject to the levels of volatility based on any sort of thing. So you can always start by my existing customer side is going to produce future revenue, and you can look at cohort specific LTV data and figure out exactly how much future revenue. That's the foundation of your forecast. Then the next layer is owned audiences. If I think about like organic SEO, my keyword rank of position two on keyword, CrossFit sports bra is going to produce for me a certain volume of traffic that will lead to revenue.

Taylor:

That's more predictable, even than advertising CPMs. The same thing with my remarketing traffic is going to be more predictable than my prospecting traffic. My email database that I've collected ahead of the moment that aren't yet customers, but they're people that I can speak to for free is subject to less volatility than those environments. The more that ahead of this Black Friday, Cyber Monday, you can develop owned audiences of communication that allow you to forecast more accurately your future revenue, the better position you're going to be in, and the less subject to that crazy volatility you're going to be. The end tactic there, out of that sort of ideology, is that right now you should be accumulating as many emails and YouTube subscribers and Instagram followers and website visitors that you possibly can to help you build a foundation for that future revenue in a way that's less volatile.

Stephanie:

Oh, I love that. Are there any specific examples of creative campaigns that you're working on with your brands right now, or that you know of?

Taylor:

Yeah. So, one of the big things that we love as a vehicle for this is quiz funnels. A lot of our brands are running these quiz funnels, and what I mean by that is you come to the website and there's like, hey, let me ask you some questions to help make a perfect product recommendation to you. And it's email gated. So, you're saying, hey, give us your email, and at the end of this, we'll send you a specific product recommendation. So, what that does, what we've been able to see is that, with many of these quiz funnels, you're able to prospect at virtually the same return as your general campaigns, but you're collecting emails at four to five to six X the rate. In those examples, what you're doing is you're almost able to match your present value, but you're banking a lot more future revenue off of all of those owned audience emails that you're capturing.

Taylor:

That's one example of the way that we think about this, of like, not just like the singularity of how much revenue I'm making now, but what additional assets are being accumulated to drive future value. Another example I would use, and this is something we talked a lot about last year, is that website traffic. So, when you go to do remarketing, which most of your revenue in the holiday period, specifically Black Friday Cyber Monday, is going to be driven out of your remarketing audiences. What that means is that website traffic today has a high future value. The more website traffic you can generate now, the better. Let's say you have two Facebook ad campaigns, and one has a 50 cent CPC, and one has a dollar CPC, but the ROAS is the same, I'm going to give preferential treatment in scale to the campaign with the lower CPC, even if the ROAS is the same or slightly less, because that traffic is going to yield future revenue in a way that's really important. Those kinds of thoughts around, what is the incremental value of the audience creation, is a really important consideration.

Stephanie:

Oh, I like that. How often are you checking in with those ... when you're starting to do that prospecting, in a way, early, how often are you checking back with them so you stay top of mind?

Taylor:

The window that I think about, so we have our general campaign set up, which is that we look at our time lag report, which in Google is basically the average window that a customer will make a purchase in. You can look at and see, like, what percentage of your customers make a purchase within 24 hours, within 48 hours? Usually, almost 80% of your are making a purchase within seven days. For the initial consideration, our primary remarketing funnels usually attempt to encapsulate the primary consideration phase, which is usually somewhere around seven to 10 days. But in terms of this strategy, what we're doing is we're building an audience that, ultimately, when it comes to Black Friday, Cyber Monday, the widest Facebook audience you can look at is 180 days.

Taylor:

The second you get within six months, which just happened for us a couple of weeks ago, we got within that six month window of this. Now, you've got an audience that you're going to speak to again at Black Friday. I don't know that we have an intentional strategy to speak to them again at a 45 day window or a 90 day window. We don't. But I think that one, that's a good idea. I think that it might be worthwhile to just sort of warm them up again, but as long as they're within that 180 day window, I have the capacity to speak to them at Black Friday, Cyber Monday.

Stephanie:

Got it. That makes sense. So, what kind of mistakes do you see companies making when it comes to these growth strategies?

Taylor:

Oh man, that's a broad question. One of the primary ones is what we call the single account ROAS problem, which is this idea that most people are running their ad accounts on the idea that every purchase is the same just based on the cost to acquire the customer. Let's say I have a target of a two to one ROAS, which basically means that I'm paying 50% of the purchase price in acquisition. Well, the reality is that not all customers are equal to you, and not all product sales net you the same amount of dollars in your pocket. One of the things that we work really hard to get really clear with our clients on is the value of customers by different cohorts, so by first product purchases an example. If you think about every skew that you have in your store, every skew has a different margin and has a different return rate and has a different shipping cost.

Taylor:

Each of these three variables mean that, even at a fixed ROAS, so if you applied a two to one ROAS across every skew that you have, the net dollars in your pocket is different for every skew. Really understanding which products net you the most dollars as a business owner is a critical data element that I do not see enough people consider, and they don't design their ad accounts to reflect the variable value of the product.

Stephanie:

That's a good one. I haven't heard anyone talk about that as a mistake yet, so that must mean a lot of companies are making it.

Taylor:

Yeah. It's because it's hard. The information is hard to access, and it's hard to get granular visibility too. It's a much simpler decision-making mechanism to go, "Well, my blended cogs are 30%, and so if I get a two to one I'm making 20% generally across my business." That's a much easier consideration to make than to try and actually break it down by individual skew and get really specific on the decisions.

Stephanie:

Yep. Are there any best practices when it comes to developing spreadsheets or dashboards or something that can give you easy access to that information that you've seen your brands or yourself develop?

Taylor:

Yeah. We've built a tool to do this, to calculate cohort specific LTV data, to ultimately give us a view that we care a lot about, which is your 60 day LTV by specific cohorts. The reason we look at that time window is because most early stage Ecommerce businesses can't really wait much longer than 60 days to really realize the value. They just don't have the cash position to be able to wait longer than that, nor do I think it's really wise to wait longer than that. We've built something internally to do it, so I wish I could say like, hey, just go copy this spreadsheet and set it up, but it is difficult information to access, but you can do it. You can build cohort models for your business, but beyond just the LTV, the easiest thing to do is to really deeply understand that you did economics by every skew that you have.

Taylor:

What I would do to start is I would export every product that you have, I would mark the MSRP, so the retail sales price. Then next to that, I would put the cost to you as a business. Then I would put the return rate of the product, and then I would put the shipping cost of the product, and then I would calculate the net value of every skew. Just being aware of that for your merchandise set, like for the entirety of your product catalog will give you the kind of information that you need to then think more specifically about your ad account.

Stephanie:

That's great. To shift a bit into, we were just talking about how to grow a company, then track metrics and all that. I know you also have just been recently talking about an anti-fragile Ecommerce playbook. I was hoping you could kind of detail what that means, how are you setting up the anti-fragile Ecommerce companies, and what does that playbook look like?

Taylor:

Yeah. This goes back to this idea that we are bad at predicting the future. When I think about how I want to build a business, I want to build a business that thrives when I'm wrong. I actually want to accept the fact that I am not going to be able to determine the future, and I want to set up the business to be able to survive in almost any environment. That's sort of the idea of anti-fragile, is not just that I'm resistant to negative, but actually the negatives can be in an environment in which I succeed. The way in which I think that you're afforded the opportunity to do that is by having a specific set of attributes related to your business that allow you to sort of thrive in variability. Some of those are ... now you're going to make me try and quote my exact tweet to see if I can remember all of them, but one of them is high margin, right?

Taylor:

This is seemingly obvious, but this idea, the more room for error that you have to still be profitable, the more that the variability on your CAC, as an example, still does it affect you, and it allows you to potentially win when others can't. So, high margin is one. Another is, and this is a really underrated one, is great payment terms with your manufacturing supplier. One of the biggest things that destroys cashflow for Ecommerce businesses is obligation to front the cash for inventory. One of the best skills I think an entrepreneur can have is relational development and negotiation skills with their supplier to develop trust, to get to net 30 terms on delivery, where you're actually realizing the revenue of your product before you have to pay for it. That allows your cash conversion cycle to speed up tremendously in a way that's super, super helpful. Another one [crosstalk 00:21:19].

Stephanie:

When I saw that, I'm like, that's huge. We have the same thing in media. People will start coming and be like, "Oh, we have net 60 or net 90 payment terms." You just can't take no for an answer sometimes, and just keep trying to build up the relationship and be like, "We really can't do that." Oftentimes, you'll be able to get down to net 30, just like you're mentioning here. So I thought that one was a really good point. Sorry, keep going.

Taylor:

Yeah, that's exactly right. You just got to think about your cash conversion cycle and how you're going to not be in a position where you have to sort or seek outside capitalization to fund when you're winning. That happens in Ecom because of the cost of inventory upfront, which makes it a complex cash management business. Another one, and again, these seem obvious, but I think we don't consider them enough is low OPEX as a percentage of your revenue. We have this principle, we call the four quarter accounting principle, which gives you sort of a directional guide to where you want your revenue to go. So, if you think about your top line revenues of business, we break it into four quarters. The idea is, if you want 25% profit, then your costs need to exist in these quadrants.

Taylor:

You need to have 25% CAC, so a four to one, what we would call marketing efficiency rating, so total revenue divided by total sales is 25%. 25% cost of delivery, which is basically the cost of the product from wherever you're manufacturing it all the way to the customer. Then 25% of OPEX, and that's your payroll, your rent and all of those things. If you look at your P&L and you look at it relative to each of those four quarters, you can figure out where your profit gets eaten up and where you need to go make improvements. So, the lower your OPEX is, so you can do this by not having a really expensive office by controlling your overhead in an employee count in smart ways. Really, just looking at how do I run this business as clean as necessary or as possible in order to give myself opportunity again, because there's going to be variability in my other places.

Taylor:

The shipping thing is a good example, where it's like, maybe my cost of delivery is normally 23%, but because of the shipping thing, maybe it's going to go to 27% of my revenue. If that destroys your profit, because your OPEX is too high, then it gives you a lack of options. That's a really important one to consider. Gosh, what else did I say in there?

Stephanie:

You also said diverse traffic mix, which I like.

Taylor:

Okay. Yeah, so this goes back to the point that we're talking about. Just like an investment portfolio, if you're over-indexed on any single channel, if that channel deteriorates in value, your business is in real trouble, but if you have a diverse traffic sourcing, and I think, so the question is, what's a diverse traffic sourcing? A good baseline metric is 50% paid, 50% organic. You're going to be able to survive volatility in any one of those channels, because you have a good amount of traffic from other sources. All of these things, make it so that when the inevitable problem strikes, your business is set up for it.

Taylor:

I think that we under consider how important it is to get into these positions of strong foundations of anti-fragility before we pursue further growth. Especially in this crazy environment where we're in now, where basically every forecast that I see every business make is wrong. The question is, what do you do? A lot of people want to try and think about like, well, how do I forecast better? I just go like, I think that's a fool's errand. I think attempting to predict the future, there are just too many inputs to do it well. So, instead, how do you build a business that when you're wrong, you still win?

Stephanie:

Yep. Yeah, I love that. I think it's such a bad mindset to think that like, you have to be perfect. I think that that's how companies do it. Even when I worked at other finance groups within companies, we were usually not right, which is why we did scenarios because you usually are wrong.

Taylor:

Well, that's exactly it. The quote that I always come back to is that, all models are wrong, some are useful. They're useful in their ability to understand where you're importantly wrong. When you have a detailed model, let's say I have a prediction that shows me what I'm expecting my traffic to be by every channel next month. Okay, so for direct organic search paid, search paid social, I have a prediction. Well, the reason that model is so important is not because it's going to be right, but because it allows me, as I'm actualizing the data, to understand where I am importantly wrong. As the data's coming in, I can start to see, Ooh, I'm missing my prediction in email by a lot, and it allows you to then think about strategies to go and solve that problem, where if you don't have that model, if you don't have that prediction, it's really hard to determine where the problem comes from.

Stephanie:

Yep, exactly. It's like the pandemic too, who would have predicted that? It's straight out of Missing, Taleb's book. That's a black swan event, you never would have predicted it, so why try? Just different scenarios, and a worst case scenario like now.

Taylor:

It seems a lot like I'm reading Antifragile as I'm thinking about the application of Ecommerce. It's like, well, how do you think about a business like that? What I see is I see frustration from our entrepreneur partners about their forecast being wrong, and they get really upset about it. I get it. It's really hard. You have to make decisions about this. So, it is an important exercise and you want to reduce the margin of error as much as possible going forward, but you have to begin to expect that. That's one of the things I think about being a more seasoned entrepreneur who have seen thousands and thousands of forecasts be wrong is I'm no longer surprised when they are. So, it just gives me a different frame of the problem.

Stephanie:

Yup. When people are trying to think of those black swan events or scenarios, how can someone go about building a scenario if they don't even really know how to anticipate it? They don't even know what to prepare for.

Taylor:

I like the idea of, rather than trying to predict the future, you should extrapolate the present. That's the Nassim Taleb idea, which is that, where are you now, and where do you believe you will be in the next month based on the present? If you extrapolate the present versus predict the future, then what it allows you to do is to think about, okay, my organic search is currently at 20,000 visits a month, and it has grown by 5% a month. If that continues, I'll be at 21,000 visitors next month. You put that into place, and then as you actualize it in real time, then you can understand what's happening. What you need to understand is that, the further out you go, the wider the margin of error becomes. Predicting tomorrow is a lot easier than predicting a week from now, which is a lot easier than predicting a year from now, because the number of inputs and variables just increase as you move out.

Taylor:

That idea of constantly re-forecasting and constantly actualizing your prediction and making adjustments, that's the skill, that's the exercise to continually get in the habit of doing and understanding where you were wrong. Then, doing your best to understand why, I think sorting out causality can often be sort of this thing that we chase and we make assumptions around. I think it's sometimes useful, but more importantly, it allows you to make adjustments in your next forecast, and then do it again and then do it again and then do it again. But again, no matter how many times you do it, you're always going to be wrong, and that wrongness is the thing that I think is really informative, and it allows you to ask the question, how do I make sure I'm okay when I'm wrong? That becomes the important thing to then go build.

Stephanie:

Yup. I love that. Yeah, it reminds me back in my Fannie Mae days, I used to do with a housing forecast, and we would literally be forecasting for like the next month, and we would have data almost halfway through that month and we would still be wrong. I think it boggled my mind how we'd be wrong, but even thinking of that, I'm like, there's just no way, once you get past a certain point, you just have to keep re forecasting, even if you're halfway through the month sometimes again.

Taylor:

This is such an important understanding about how we as humans process information. My favorite example is something I call the roulette run problem, where if you've ever been to Vegas, there is a reason that at the roulette table, they display all of the recent numbers for you. The reason is, is because what your brain does, when you see a bunch of reds in a row is you go, uh, well, the next one has to be black. And you build this relationship between past data and future data that is not real. We do this all the time with information as humans. We're just really bad at computating things in a statistically rigorous way about the future. This is also why humans are actually really bad media buyers, and why we try and train all of our internal media buyers to make as few of decisions as possible inside of the ad account.

Taylor:

Because if you think about what most people are doing inside of a Facebook ad account, is they're loading up a dashboard. They're looking at past data. So, they're looking at historical data and making inferences about the future without using a computer, without using a calculator, without so much as writing down chicken scratch. They're trying to make predictive decisions about how things like CTR and CPC and ROAS are going to relate to the future, and they're almost always wrong in those decisions. This is just like, it's really important to understand what the biases are that affect us as humans in our decision-making.

Stephanie:

That's great. Talk a little bit more about humans being bad media buyers, how are you all going about buying media? What are some best practices? Other than just saying like, "Okay, just rely on the platform, let it do its job." What kind of things are you guys trying out and doing and seeing success with right now?

Taylor:

What I'm going to push back on is that language that you use, I think diminishes the right idea, which is people love to say like, "Oh, well, yeah, just let the algorithm do it." They say that in a way that reflects that that's the simplistic decision. What I'll tell you is it's not. It actually takes incredible discipline to be thoughtful about allowing the computer to do what the computer does best and focus on what you do best. With our media buyers, their job, if you think about any machine learning tool, the key to a great machine learning tool is the inputs. Machine learning tool is just taking a set of inputs and understanding them to generate future outputs. Well, the key to being a great media buyer is you set a good structure of inputs. I'll give you an example.

Taylor:

Let's go back to this idea of different product values and different average order values by purchase types. If you were to export your last 30 days of orders and you were to build a scatter plot, where across every day, every dot was an order by AOV, you would have a variation of different order values. If you think about one of the most common mistakes I see in an ad account is you have a campaign that's bidding for lowest cost, which is the conversion objective, where you're basically saying to Facebook, give me the lowest CPA that you can inside of this ad set. Then in that ad set, what they will do is they will sell a bunch of different products with different prices. Let's imagine I have five different skews. I'm a jewelry brand, and they range in price from $50 to $500. Well, what product is Facebook going to favor if the structural setup is lowest cost and you give it those options?

Stephanie:

Yeah, the cheapest one.

Taylor:

So, you're building a structure where the computer output is going to be focused on the skew that likely doesn't generate you the best net outcome. That is a tactical media buying error that has to do with poor structural setup and understanding of the tool, that has nothing to do with the decision making of which ad you should allocate the budget to. That's the kind of thing we teach our people to think about over and over and over again is, am I setting up the right structure for the output that I want?

Stephanie:

That makes sense. You would instead maybe have like similar price products so they can all actually see the performance, instead of teaching the algorithm to, of course, always showcase the cheapest one because you don't have a budget?

Taylor:

That's right. Or I'm going to bid for highest value instead of lowest cost. Facebook has different conversion objectives I can set up relative to the thing that I want. The question is, and we play this game at our company every Friday morning called [inaudible 00:33:36], where I pull up an ad account and we go through campaigns and I make the media buyers tell me, what are you intending to accomplish with this campaign? What were you trying to do? Then let's discuss of whether the structural setup and inputs were right for that outcome. Did you design the system accurately to generate the thing that you want? Because Facebook's tool is incredibly good at getting you what you want, but over and over and over again, I see people design systems that will never, ever generate the outcome they want. So they're setting up games where the rules mean they will never win.

Stephanie:

Yup. That's really good lessons. Any other things that you've seen, like similar themes of ad buying, where you were like, I've seen quite a few people do it this way and it's wrong. Anything else that has come on your radar?

Taylor:

Yeah. The big other thing is just the relationship of budget to number of variables. One of the things, again, this all comes back to some basic statistical ideas, where if you think about ... have you ever seen a graph of how long it takes to normalize the data for flipping a coin? Like how many coin flips it takes before you basically get to 50% outcomes? Have you ever seen a graph that looks like that? What it is, is that like, when it's a two sided coin, the amount of flips that it takes to normalize the data, so in other words, the time at which it will reach the predicted outcome of 50%, it's somewhere, usually in the range, by the time you get to 100 flips, it's almost always going to be at 50%. Now, if you take a six sided die, okay, do you think it takes more or less rolls to get there?

Taylor:

It takes substantially more, because there's more possible outcomes, so the amount of time that it takes for the data to normalize is a lot more rolls than if there's only two possible outcomes.

Taylor:

So the same is true of the way that you build a campaign. The more end nodes in your structure, so think of an end node as an ad set or a campaign to an ad set to an ad. The number of variables in your campaign is going to determine the amount of budget that is required to get to statistical significance, to get to accurate outcomes. The other big mistake I see people make is they build these campaigns with all these end nodes, all of these ad sets, all of these ads and a very tiny budget. You might as well check back in three years and then you'll actually get a definitive result. What they do is they build campaigns like that, and then they make decisions on insufficient data. What that is, is it's basically looking at four flips of a coin and then saying, "Oh, it's 70, 30 heads. The truth here must be this coin yield 70% heads."

Taylor:

Instead, you have to build structures that allow you to get to accurate information quickly by having an appropriate amount of budget against the number of nodes in your test, if that makes.

Stephanie:

Yeah, that makes complete sense. Is there any ratio where you're like, for every 10 ads, you need $1,000, or is there any ranges like that?

Taylor:

Facebook gives you this information. They tell you that they want 50 conversions per ad set per week. That's what they need to get you out of the learning phase. There's a lot of people that are these Facebook truthers, that think like Facebook has all these ulterior motives to attempt to get you to spend money. I get it. They are rationally self-interested, but your success is actually in their interest. So, they will give you directional understanding of how to use their tool best. So, part of what they tell you is that they designed this thing called the learning phase, which is basically their way of declaring that the data that you're seeing is not trustworthy yet. When you are still learning, you should not act on this data because it's not actually predictive of the future.

Taylor:

But then when you get out of learning phase, now you're in an actual set of outcomes that are more predictive, they're more accurate, they're more deep, they're more true about the set of inputs that are there. So, the way to get out of that is this idea of ensuring that, based on your budget and based on your target CPA, you can get to 50 conversions per week per ad set. Because campaigns use a daily budget, if you take 50 and you divide it by seven, the formula that we give people is 7.14, which is just 50 divided by seven, so for your daily budget. 7.14 times your target CPA. Again, that's the payment that you want for the objective times the number of ad sets in the campaign, that needs to be your daily budget in order to get through optimization as fast as possible.

Stephanie:

Ooh, that's good. I'm writing down the formula now, so we can put it in our show notes.

Taylor:

Yup.

Stephanie:

Okay, cool. Awesome. That's really good tips around Facebook ads. I agree about the, like when you were mentioning the learning phase, we have our own ad network, and it's just the same thing. Our growth team's always like, "Hey, we need about $1,000 just to learn and then we'll let you know what the cost per click is." It's not just Facebook, it's other platforms as well.

Taylor:

That's right. People get frustrated by that because it's a media buyer, especially if you're a company that's charging on a percentage of ad spend, it's really hard for that to feel like anything, but a self serving piece of information. But the reality is, it just goes back to an understanding of how data happens. It's like, again, if you wanted me to tell you the results of flipping a coin, give me a hundred flips, not five, and it's the same thing. That's the understanding that we need to get to about how you get to statistical significance.

Stephanie:

Yep. So, are there any surprising campaigns that your team has initiated or certain kinds of ads that you were like, "That's not going to work", and they actually performed well?

Taylor:

Oh man, all the time. I'm so horrible at predicting what will work well, creatively in particular. Man, anything that I think is like a really interesting format right now.

Stephanie:

I feel like the formats right now are ... they have to be kind of different. There's a lot happening with the ad platforms because of brands pulling back and other brands dipping in on those ad platforms, but also just with the styles of ads that are going out. With the whole world, it just seems like things are different. So, I'm wondering how you guys are approaching that.

Taylor:

Yeah. Different is a good thing. The hardest thing is actually to avoid ruts. Because there's this horrible habit that we have in the media buying world. I'll lump us into this. We fall guilty to it too, which is the way that you generate ad creative is you go to Facebook ad library, you look up your favorite brand and you copy the styling of their ads, because you're making an assumption that they're working. But I'm a big believer in sort of the purple cow principle from Seth Godin, which is this idea that, the value of an ad deteriorates in repetition of its use. Every time that an ad goes out into the world and every time it's replicated, it becomes less valuable. That same thing happens with ad formats. We've seen it with influencer ads and UGC and the Mashable style, and these formats that have become really popular, eventually their impact is reduced as people encounter them.

Taylor:

I think that the key thing, and this is the biggest challenge inside of an organization like ours, is how do you produce a system that constantly generates novel ideas? That's what I would think a lot about. Again, I know this isn't the simple answer, where I can say to you, just use an ID story squared, whatever format, but I would be lying to you if I said that. Instead, I would really take space to challenge yourself to think like, one think a lot about the beginning of the ad. I know this is cliche, but it's just so true, is that video average watch times on Facebook are like four seconds for great ads. You just don't have as much time to say the thing that you're intending to say. It's sort of the David Ogilvy quote, "If you're selling a fire extinguisher, lead with the fire." That is just fundamentally true.

Taylor:

I would think about that as a core principle, but beyond that, you've got to break the feed. You've got to be novel, and you've got to figure out a way to differentiate yourself. People are scrolling a mile a day on their phone. If you plan to stop them and break them, then you've got to figure out a way to be compelling right off the bat, and a great metric for tracking this, so we have this principle that we sort of developed off of AIDA, which is a sort of legendary advertising hierarchy of effects model, and we've sort of applied modern metrics to it. So, if you think about the first one, the most important one before ... so with the hierarchy of effects model, they happen in sequence. The second one can't happen until the first one happens. I think this is really true of Facebook ads, which is you don't get a chance to communicate your message until you stop them and capture their attention.

Taylor:

A metric we focus on a ton is three second views divided by impressions. In other words, what percentage of your served ad impressions result in a three-second consumption? You want to see that number get close to a north of 40%. Otherwise, you are paying for a lot of people to scroll past your ad.

Stephanie:

Yep. I like that. I like having specific metrics set up for the AIDA format, which for anyone who's like, "What's AIDA?" Attention, interest, desire, action retailers.

Taylor:

That's right. That's exactly right. We assign a metric to each of those variables, so attention is this three-second views divided by impressions. Interest is video average watch time. Desire is outbound CTR. Then action is ultimately CPA or ROAS or whatever you use there. But the important thing again, with these models, is to think about them happening, they have to happen in sequence. What that does, what I love about using that as a feedback loop for our creatives is that the worst feedback that you can give to a designer or a creative is your ad's not working. That is so not helpful, because what do you do with feedback that says it's not working? What you can do is if, instead of thinking about the ad as a single unit, I think about it as component parts. If I say to them, "Hey, your three-second view to impression ratio is really low."

Taylor:

Well, now I can think about what's happening at the beginning of the ad. Or if I say, "Hey, everybody's dropping off at four seconds." Now, I can think about what's happening at that point in the ad. So, you can start to now deconstruct the actual ad into pieces that allows a creative to actually take it and iterate in ways that are productive.

Stephanie:

Yeah, I love that. It seems it could get a bit overwhelming if you have tons of ads running, like starting to just try and dissect, what's happening at the four second mark, or what's happening when it should be the D in AIDA? How are you thinking about, if the company has like a large amount of ads that they're testing?

Taylor:

This is a great question, because this is the number one marketer's dilemma, is that, what do I do next? There's a thousand million gazillion things to do. The answer is you sequence by volume of opportunities. I'm going to just start with the ads that are running in my campaign with the highest spend, and I'm going to iterate on those first because that's my area of highest potential impact. Then I'm going to sequence the rate at which I engage with my ad creative relative to their potential impact. That's a really important thing to think about is that, the sequence by which you decide to do things has opportunity cost to it. So, you have to make sure that you are going after the areas of highest impact relentlessly first.

Stephanie:

I love that. Yeah, really a good way to think about it. How are you thinking about maybe ... oftentimes, history repeats itself, how are you thinking about looking at historical ad styles or going back to direct mail, which a lot of people are doing right now since everyone's at home? How are you thinking about viewing history to maybe impact present day advertising methods or growth strategies for your brands?

Taylor:

Yeah. We have this course that we teach to all of our designers called advertising philosophy. The idea is to do that exact thing, which is to understand the fundamentals of advertising that are always true. So, we start with things like behavioral psychology. We read The Choice Factory by Richard Shotton, which talks about different human biases, to help them understand things that are just true of people, period. Then from there, we sort of begin to allocate that against the medium. We read a little bit on advertising and the history of advertising the change, the history of advertising, which are all just chockfull of amazing adding examples that are primarily prints from a legacy of magazine print ads, but they're so much in those print ads. If you think about how hard it was to actually create impact with a print ad, like someone had to open a magazine, they had to see the ad, then they'd to physically go somewhere to take action. It was way harder to be an advertiser there.

Taylor:

So, the copy had to be so strong and the visuals had to be compelling. There's so much to be learned from a lot of those principles that I think we underestimate the value of the art and history of advertising before we worry about applying it to the modern medium. So, we teach that to all of our people, because one, we want them to fall in love with advertising, because we think it's an incredible art form. A lot of times, for creatives, advertising is sort of seen as the thing that you do when you fail at being a real creative. We just don't believe that. We believe that it is like a true, true art form that reaches millions and millions of people, and if you can learn to really love it and to love the way in which you can impact human behavior with your communication, it can unlock like inspiration for you. That's a long winded way of answering, and it's not just as simple as direct mail, but you've probably realized by now my answers are rarely simple.

Stephanie:

No, I love them all. Every time I'm Like, ooh, that's a good quote. We should pull that one, because yeah, that's a really good point. Even thinking about when we're generating ads for Mission, it takes so much brain power to think about like, what's good copy, what kind of images should we use? What is someone going to remark about? It's a process. It's harder than even building out a podcast sometimes. So, I can [crosstalk 00:47:18].

Taylor:

Yeah, as a simple hack, because I think these things are cyclical. If you are really creatively stalled and you're looking for inspiration, I would really encourage you not to use Facebook ad library as a mechanism for copying. Go look at one of these books, they're full of ads, and literally copy an ad from 1960, and think about the language, the imagery, and literally replicate it in the modern day. It will be more impactful than copying somebody's ad from Facebook ad library. That is a much better source of really potentially impactful ad creative than is the stuff that you're going to see from basically every brand on the internet.

Stephanie:

Yeah. That's really great. Are there any brands that you're watching that are actually always head of someone, where you're like, these brands all have the same ads, but this particular one is always kind of ahead of the game?

Taylor:

Yeah. I put out the other day, like one of the things I've gotten shamelessly relentless about is when I see something amazing, I go try and figure out who did it and try and hire them. So, TRUFF, the hot sauce company, their ad creative is as good as anybody's I've seen in the last few months. I put out on Twitter the other day, like I was citing that as an example. I just said, whoever did this, if you reach out to me, I'll double your salary to come work for me right now.

Taylor:

I think what they've done is really, really fascinating, clever ways of taking a sort of historically boring product in hot sauce and making it super compelling. One of the things, a lot of creatives feel constrained by the product attributes, like our initial entry point in ad creative all the time is like the product features. We just have this phrase that we tell our creatives all the time, which is you live in an infinite creative universe. You can say and do anything and make it relevant to this product. One of my favorite examples of this is there's this legendary ad from Gillette razors. It's a picture of an egg in a frying pan, and the headline is, this is an ad about razorblades.

Taylor:

The copy just talks about how the nonstick material in the pan is the same thing that they use to build the razors so it doesn't grab your face when it goes by. It's just brilliant. But the idea is like, it's a picture of an egg in a frying pan. It didn't come with the asset library from the client when they sent it over. But if you open yourself up to the possibility of telling stories broader than just the library of photos you receive from the client, and you really embrace this idea of an infinite creative universe, you can do anything. I think that's when you get to the really interesting stuff.

Stephanie:

Oh, that's great. I love that. Yeah, one of the ads that a friend's company is running, it's for socks, but it's literally a piece of corn growing out of two different. Something like very random, not relevant and it's their best performing ad.

Taylor:

That is the thing. It's like random. Again, you've got to think about somebody who is scrolling endlessly on their phone, and what is going to stop them

Taylor:

It's just about understanding, and this is the thing. You've got to be an internet user. In the same way, like everyone thinks about platforms like Reddit is this really specific sub-culture where you'd have this specific language. I saw a hilarious ad for a Reddit out the other day that understood this really well. It was Fresh Box, or one of the meal delivery services. The ad was basically like, "Hey, Reddit, my boss asked me to run an ad on Reddit, so I went ahead and took care of it for you." It's basically this meme app that Reddit users use, which is a robot blowing up the brand. It was basically the satirical way of understanding the medium so clearly, that they mock themselves in a way that made them endeared and loved by the Reddit users. That sort of ability to understand the place and the environment in which people are receiving your content is a real skill.

Stephanie:

Yeah. No, that's a really good example. I know we only have a couple minutes left, so I did want to ask a higher level Ecommerce question because from the interview, I just know you're so excited about predicting the future and you just feel very confident about it.

Taylor:

Do it.

Stephanie:

I wanted to hear, what kind of trends or patterns or disruptions are you most excited about right now in Ecommerce, or do you see coming down the pike?

Taylor:

Let's see. One of the things that I like, I'm sort of betting on for the future, is the global Ecommerce market. I think about the ... when anytime I see a marketplace where there is this pent up demand, that the infrastructure doesn't yet support, but it's finding a way to happen even before the infrastructure exists, what that means is that, by the time the infrastructure catches up, you're going to have a massive moment of arbitrage. Think of it like a wave that builds up that suddenly then released by some sort of technological innovation, and that's where there's periods of arbitrage before everybody else shows up and the competition's there. I think about a market like India, where you have this massive, massive market that is coming online really, really fast, and you have this problem, which is a payments infrastructure.

Taylor:

Nobody there has credit cards. Almost everything is cash on delivery, and none of the systems yet support the infrastructure for that delivery and payments processing. But yet, there's massive engagement from the user base. The second that gets solved, it's going to be a huge opportunity. This is true in a ton of other countries, where you have ad inventory costs that are extremely low, you have demand that's really high and you just don't have the systems that need it yet. That's a thing I'm watching and following really closely, as we already have clients that are seeing 30, 40, in some cases, 50% of their spend in international markets, where there's tons of friction in fulfilling the product. The second that, that all goes away, we are on the precipice of a truly global marketplace in a way that we haven't even begun to process. It's going to be really, really cool.

Stephanie:

Ooh, that is a very good answer. What other parts, or what other places in the globe are you looking at right now? India of course, is a big one next million users, everyone's focused there, but what other smaller markets are you looking at right now that you see a big opportunity, or your brands [crosstalk 00:56:20]?

Taylor:

Places where like, so the United Arab Emirates is anytime you ... especially, if you have any sort of luxury product, there's massive opportunity there. There's highly efficient opportunity. It's pretty volume constrained. It's a small country. Super interesting. I think Southeast Asia is super, super interesting to me. Nigeria, even from a talent standpoint in the Ecommerce world, that's a giant country. I think sometimes, because we have these geographically warped maps because of our Western centric view, we don't realize how big some of these countries are. Those are a few, Nigeria, Southeast Asia and India from a hiring standpoint, from a potential market opportunity standpoint are places that we're following really closely.

Stephanie:

And is good spots to watch. All right, so we have a few minutes left. I want to hop into a quick lightning round brought to you by our friends at Salesforce Commerce Cloud. This is where I'm going to ask you a question and you have a minute or less to answer. Are you ready, Taylor?

Taylor:

Ready, fire.

Stephanie:

All right. What's next on your reading list?

Taylor:

What is next? I'm going to pull up audible right now. Do I have a countdown on my clock? Shoot. Where's my reading app? Audible. Someone just sent me a book. Oh, I know what it ... Here, let me take it. I'm failing at the time challenge, but [crosstalk 00:57:36].

Stephanie:

That's okay. Thankfully, we can edit this podcast and make it [crosstalk 00:57:39].

Taylor:

Yeah, a book called Caste: The Origins of Our Discontents, by Isabel Wilkerson. It's all about sort of racial history in the United States and some of the things that have led to our own version of a caste system. This is a big sort of initiative for me personally, in my own learning and development right now, is sort of redesigning my own sense of history, as well as we've got a bunch of big diversity inclusion initiatives inside of CTC right now. So, I'm trying to do the work of my own education, and that's a book that was just recommended to me.

Stephanie:

That sounds good. I will have to check that out as well. What's up next on your Netflix queue?

Taylor:

Man, Netflix is so dead right now. It's one of the biggest drawbacks of COVID era. We just finished Money Heist, which is awesome, or Heist, maybe it's just called, which is dub from a Spanish show. So good. But I'm watching a show on Amazon right now called Upload, which is about this idea that like in the future, they've invented a way that when you die, they upload your consciousness into this digital world and you can interact with the present world. It's sort of a comedy, but it's ... I like those sort of futurist dystopian content.

Stephanie:

That's interesting. If you were to create a podcast, what would it be about? It can't be Ecommerce, and who would your first guest be?

Taylor:

Right now it'd be about trading cards. I've become obsessed with baseball cards in the last few months. My guest would be God to the name of Evan Vandenberg, who's launching this company that I might potentially be an investor in called [CollectX 00:59:07], which is basically, think of it as Robinhood meets DraftKings, but it's basically the digitization and tokenization of physical cards into a market dynamic that I'm sort of obsessed with. That's a weird nerdy answer. I'm sorry.

Stephanie:

I know nothing about trading cards, but that sounds very intriguing. All right, and the last one, what favorite piece of tech or an Ecommerce tool are you trying out right now that is either making you more efficient or you're having a lot of success with?

Taylor:

It's one that we're building, which is our growth data tool, which is just all around cohort specific LTV data. I have this super fun position of being able to hire a rad developer and a rad product manager and just build something that I want.that is just so fun. To have ideas in your head and then for them to be able to manifest themselves into the world, it's magic. That's my real answer, but I'm an obsessive Evernote user. I believe a lot in creating an external hard-drive for your brain as a really, really powerful tool. I think we were already sort of bionic people more than we realize, and expanding your mental capacity by taking great notes, I think is a serious life hack.

Stephanie:

Yeah. I completely agree there. All right. Well, this has been an awesome interview, so many good insights and tips and things that people can actually implement, which I love interviews like that. Where can people find out more about Common Thread Collective and yourself?

Taylor:

So, commonthreadco.com is our website. Then, I would say follow me, Taylor Holiday on Twitter.

Stephanie:

You are a good Twitter follow. Yes. All right, thanks Taylor.

Taylor:

Awesome.