Latest / Frankly eCommerce / What AI Didn't Solve about Data Analytics

Transcript

What AI Didn't Solve about Data Analytics · Frankly eCommerce

• 0:00
  There's no shortage of people wondering how the return on ad spend is gonna play out this year You don't need a back and forth. You don't need to second-guess yourself. Watch this KPI. Here's where it can reach.
• 0:10
  As long as it's-- When it reaches that, just dial back spend and go ahead and spend your time on this bigger opportunity out there. Where's the profit in all the ad spend?
• 0:20
  Be very careful about letting your marketing agency grade their own homework. I see it happen all too often where the agency spends your money, and that same agency says, "Here's how we did spending your money."
• 0:35
  And it's not hard to imagine that almost always they, they're giving themselves very, very high grades. Said, "I'm nervous about spending more because my ROAS is going up." I said, "Is it okay if ROAS can go up?
• 0:47
  And how high can ROAS get while they still make money?" You're listening to the Frankly Ecommerce Show, candid conversations about business with host Patrick Scott Pittman.
• 1:01
  Well, as the big ad spending period of the year approaches, Black Friday beckons, there's no shortage of people wondering how the return on ad spend is gonna play out this year.
• 1:11
  The costs go up, the clicks get more expensive as the year gets closer to Christmas, and that means we need to figure out, are we making money in all of this effort? Where's the profit in all the ad spend?
• 1:23
  And there's no one better in my mind to bring to this conversation than Dan Hansfeld.
• 1:29
  He's someone who's described as a fractional head of analytics, and in my experience, he's someone who helps us bring confidence to where we're finding that incremental benefit to the next dollar in acquisition activities.
• 1:42
  So I wanna invite all the way from Colorado to here in Austin, Texas, Dan Hansfeld. Welcome to the Frankly Ecommerce Show. Thank you, Patrick. I'm happy to be here.
• 1:54
  Now, Dan, we had a conversation about why AI hasn't just solved this data analytics problem for how many brands, and I've been all year hearing about how Claude can just sort of do reports and make it all magic.
• 2:09
  One, two, three, it's skilled, it's done, it's repeated, it's scheduled. And people sort of step back from that and think, "Well, okay, phew, glad I got to take care of that problem."
• 2:19
  And as the years unfolded, you know, I, I just found myself going back to asking questions like, well, does Claude have access to all the data that it needs to make decisions?
• 2:31
  Did it have access last week, but does it have access this week? Are the reports even the right reports to be looking at?
• 2:38
  And maybe beyond that, how are the reports we used to have generated by people no longer giving that sense of thoughtful study that someone took to create them?
• 2:50
  I don't know, all kinds of turbulence in terms of how we think about staffing for an ecommerce brand that wants to make sense of its numbers.
• 2:58
  You're in the midst of this every day, and so we had a conversation about that question of where has AI not solved the data analytics problem, particularly for the brand who's selling in ecommerce.
• 3:11
  Help me understand how you approach that question.
• 3:14
  Yeah, it's a great question, and I think the best place to start is thinking about what AI has solved, because there's a lot that AI does that is amazing in analytics and makes my job far more fun because I get to focus on the part of the work that truly drives value and that I get the most excited about.
• 3:33
  The piece of analytics which involves pulling numbers together, putting up a report, sending that report out to stakeholders, that has always been a commodity.
• 3:42
  In the past, you did need someone on the team to just press the buttons to get that working, like so many other functions in the organization. Today, Claude makes that easy.
• 3:50
  The easy question from an executive who might not be comfortable on the software but wants to know sales from last week used to come through an analyst, and the analyst, who might be me, would say, "Of course, I'd rather teach you how to pull this yourself, but here you go."
• 4:04
  The-- What we have now is you can get the insights very easily at a click, and that's ava-available to more people in the organization than ever before.
• 4:13
  The piece that's always been the true value of data, though, is not just getting to the insights. It's understanding what matters, why it matters, and most importantly, what to do with it.
• 4:25
  And that's as much a challenge today as it ever has been.
• 4:29
  And while there's still a place for AI in that piece as well, that's a piece where I'm finding brands still need somebody who understands the data and what questions to ask to get to the point of value that they're really looking for.
• 4:43
  Does that change based on the size of the company that you have? Do you feel like your advice would be different for a big company versus a small? It changes based on the people at the company.
• 4:54
  I've met, uh, individual founders who are still on their own, who have some data experience, who are handling everything on their own now, and they're doing great.
• 5:05
  In the past, they just didn't have enough time to use the expertise they had on data to go in and answer some of those trickier question-questions they have.
• 5:14
  I-- There are big companies with data teams, and now those data teams spend more time thinking through presenting data to, to investors and to the executive team, how to ask better questions, how to dig in and find insights no one had thought about looking for before because they have more time saved.
• 5:32
  However, there's a lot of teams out there across the spectrum who don't have someone on the team who has the expertise to say, "We ask this question this way.
• 5:41
  Claude might not have understood that when we said CAC, were we talking about blended? Were we talking about platform specific?"
• 5:47
  When we're thinking about our revenue and we're thinking about retention, we haven't taken into account to present it to our AI that we might be doing an initiative to increase retention in this way and releasing this new product line and expanding to this new market.
• 6:02
  All of that will change the numbers. You need those people on your team to ask those questions and to know when to ask them.
• 6:09
  And then when the data comes back, to be able to point to which metrics matter, and finally, to know which metrics actually create the story they're trying to tell in an accurate way. You use the word story, which
• 6:24
  I think is so much a part of marketing and holding attention, but there is a story to be, to be discovered in data.
• 6:33
  And so help me understand why you use that word and how, how someone should be thinking about holding in their mind this unfolding story about what data is telling them.
• 6:44
  Data on its own is sometimes, in my view, incorrectly used as the end within itself. Teams will say, "We have the data points, we have the information from the data, here we go."
• 6:57
  What I like to represent-- position teams to think about is that data is a tool that helps us get somewhere, but it's not the end result. It never is, and it never was.
• 7:08
  What data helps us do is build a story that is built on numbers, built on reality, built on confidence in a way that can be extremely helpful to marketing, extremely helpful to decision-making, operations uses it, executive teams use it.
• 7:23
  But that story that you create through data tells you a picture of your customers and your business that you otherwise wouldn't know, and that's the piece where data truly earns its keep.
• 7:36
  When you say earns it keep, I think for someone who's trying to understand how much they need their, their, their go-to-market team, does it need to have a data analyst on it anymore?
• 7:48
  Or can the tooling be set up in such a way that that becomes optional or, or even wise to outsource or, or schedule? How do you think about building up that, that team and where that fits today? Mm-hmm.
• 8:02
  You need someone on the team who understands data and who understands how the inner workings work for a few reasons. You mentioned before, sometimes the right data points aren't being pulled together.
• 8:13
  Um, other times, as I mentioned, there might be multiple definitions of a metric.
• 8:17
  Um, other times you might just have a dashboard with twenty metrics on it and be trying to figure out which one actually matters for the problem at hand.
• 8:25
  As a tool, you know, you can think of all the metrics as different tools in your tool belt. Different ones are right for different projects.
• 8:32
  Without knowing which metric is right to answer which question, you might be trying to hammer in a nail using a tape measure, um, so to say.
• 8:41
  So what I would say is, you need to be sure you have someone on your team who is confident about using it. For larger teams, that usually means there is still a role for the data analyst.
• 8:52
  Um, for the teams that I work with, my role has... I mentioned before, my work is more fun now.
• 8:57
  With the teams I work with, now I'm spending more time getting into the fun questions, the interesting questions, I should say, about what's really working for the company, which promotion should we run in the fall that's going to drive the greatest lift in sales?
• 9:12
  Our retention cohorts feel like they're down. Is that being driven by poor acquisition, or are we doing something wrong in our email or SMS campaigns that we can be improving? Those are the types of questions now.
• 9:24
  I'm not saying, I'm not saying data directly, but we wanna bring data into them, and that's now where you use the data-driven people in your company. Yeah.
• 9:35
  Tell me more then about, you also have described there are tools that can then use data to take productive actions and, and how you distinguish that from AI or just data tooling. Can you make that distinction? Yeah.
• 9:52
  Um, so it sounds like what we're getting at now is the different tools out there that present your data and bring it together. Um, there's a lot of SaaS tools out there.
• 10:02
  There's also Claude using MCPs can pull a lot of that information together. That's the piece that's always been critical to-- as a starting place.
• 10:11
  Uh, even if we have-- even as we have tooling that might directly be on Shopify or Facebook ads getting better, there still is the need, and rightly so, to bring these various data points together.
• 10:23
  Uh, data tooling is a critical piece of the data landscape.
• 10:26
  I don't think that it's as critical for as many co-companies today because you can start to rely on MCPs to pull some of this together and start to make really good decisions with it.
• 10:34
  Um, but data tooling, my biggest advice would be, before you start spending money on that, make sure you have somebody in your company who can actually do something with it.
• 10:45
  If it's simply reporting, Shopify's reporting has been getting better and better over the past two years. You might already have what you need.
• 10:52
  It might not be needed of your budget to be spending more on a, on a fancy reporting tool that moves the numbers around in a new way.
• 10:59
  You need somebody on the team who knows how to use it, um, and they have to have the time and leverage to do so. If we think about that story, is that the sort of thing that you can get from a quick dashboard?
• 11:11
  Or where does the story start to e-evolve and think about maintaining it over the course of a promotional season? We did this, this happened, then we tried this. Is that where dashboards fall short?
• 11:23
  What does the data story start to look like, and who keeps that? Great question. Dashboards are one of the critical tools for any data analyst, and dashboards are often the starting place for identifying where to dig in.
• 11:37
  Typically, dashboards won't give you the answer you're looking for. Rather, they're your quick look to spot what needs attention and to spot what might be askew and needs a deeper dive.
• 11:48
  In my workflows, typically, I'll start with a dashboard or report to get a high-level sense of where the numbers are moving.
• 11:55
  Once I've identified something that could tell me a bit more about a trend or a question, a business problem we're trying to answer, that's when I start digging deeper.
• 12:04
  It might be by working with Claude and asking questions to get deeper into it. It still might be within, you know, other data tools that are going to compile SQL, uh, manually.
• 12:14
  Um, you know, I, I use, I use Claude a lot when I can. I think it's also I'll hit that wall oftentimes that we all seem to in our area of expertise, which is when you're an expert in an area-
• 12:24
  A- AI does a pretty good job. It gets it right, right quite a bit. It saves me time. Um, it's an area we know less about, it feels like magic.
• 12:32
  So certainly when I'm working in analytics myself with Claude, I find that for the dashboarding stuff, reporting stuff, and maybe getting one step further, that's where I'm using it the most to really get into solutions and to outcomes that the rest of the company who might not be looking at data as much is now thinking of.
• 12:51
  That's where it's taking a deeper level of investigation and thinking and questions that might not be as obvious to get to. Okay.
• 13:00
  Well, to get more specific about this, one of the things that's gonna be happening in, in a few weeks is you're gonna come to Austin, and we're gonna have a workshop for brands who want to get this dialed in for the fourth quarter.
• 13:16
  And when we speak about tools or processes, these are the sort of things that we'll be able to do together in person, gathered around a big table, getting our hands, you know, our hands dirty, so to speak, in that way.
• 13:28
  But I wonder, for someone who is not maybe able to join us in Austin, how do you start to-- for this point in the year, we're going into the not yet fourth quarter, but we're getting close.
• 13:40
  Is there something that you might sort of forecast for someone to sort of say, "Let's go through a quick checklist of the things. Are you ready to spend a lot more money on ads and feel confident-" Yeah...
• 13:51
  "in your-- in why and how you're doing so?" What are the sort of things you look for, the kinda questions you ask initially? The place I start when I'm working with data is I start with: What are we trying to do?
• 14:02
  And data's gonna come in later, but the first step of the process doesn't touch your data at all. The first thing we wanna do is clearly establish a goal that we're going to be working towards.
• 14:12
  And in this time of year, it's: What is your goal for the Black Friday season? Is it a growth goal? Is-- Are you trying to hit a certain sales percentage, a growth percentage?
• 14:21
  Are you trying to bring in a certain customer? That's where you want to start. The second piece is then asking: Where can data help us reach that goal? What I want-- like to have,
• 14:32
  um, help companies avoid is starting with the data, saying, "Hey, let's dive in, see what we can find out, see if there's nuggets hiding in there," because it truly is looking for a needle in a haystack.
• 14:42
  And that's where I think companies can really be driven askew, is when they're mining the data for insights.
• 14:47
  There are plenty of use cases for data mining, but not so much in e-commerce, and certainly not when you're getting ready for a big season like Black Friday. So start with a plan. Understand your goal.
• 14:58
  From there, start to think about where data can help.
• 15:01
  Some of those are to define your objectives, exactly what you're going to do to reach your goal, and then what very specific KPIs are going to help you measure if you're achieving it.
• 15:13
  So one, if, for example, if you want to mostly be driving growth for new customers, then you need to be very careful about blended metrics like a ROAS that's combining all of your sales.
• 15:23
  You want something like a new customer cost of acquisition that's going to be laser-focused on how much it's costing you to bring in each new customer, so that you can be tracking that and using that as your ceiling for gr- for how much you push forward your ad spend, um, while remaining profitable and remaining focused on your goals.
• 15:42
  You got a signup form for your email newsletter? Of course you do. It could be better, right? There's a time and a place to make it better.
• 15:51
  But if you are not also identifying visitors to your site, de-anonymizing them, and then automatically adding them to your list in Klaviyo, your growth will be slower. How much you wanna grow this year?
• 16:04
  TIE is sponsoring this season of Frankly Ecommerce to help the brands that wanna grow faster. Do you know about TIE?
• 16:11
  When you install TIE into your e-commerce website, their network identifies the email address of many of your website visitors, then TIE adds them to your list.
• 16:23
  What you do next, when and how you send that person email, that takes care. Agencies like eBusiness Pros can help you do that, but TIE makes it possible.
• 16:34
  If you're curious, go to meettie.com/frankly and ask for a free three-month trial as a listener of Frankly Ecommerce. Now, back to the show.
• 16:47
  Now, I run a marketing agency, and we help think about these kinda things for brands. We do not focus on paid traffic and acquisition. But what if someone were to say, "Well, my ad guy, he's got that taken care of.
• 17:02
  He knows his measures. He knows how we're doing. Of course, he's optimizing to the best outcome." What is your response to that maybe common refrain? Yes.
• 17:11
  So I'm often hearing that, um, with ex- a lot of companies are using an external marketing group, and my answer is, be very careful about letting your marketing agency grade their own homework.
• 17:24
  I see it happen all too often where the agency spends your money, and that same agency says, "Here's how we did spending your money."
• 17:33
  And it's not hard to imagine that almost always they, they're giving themselves very, very high grades. That's where you need somebody using the data in a responsible way internally to double-check the work.
• 17:45
  Um, I-- There's a lot of agencies doing good work out there.
• 17:49
  There are plenty of times when an agency is, is, is spending money in a way that can be improved, um, or is spending money for a company where it just, it, it's not working.
• 17:59
  D- Having internal check, it's critical because it's really difficult for someone to say, "We spend all this money. It's just not working." It might be that your target market just doesn't exist on this platform.
• 18:11
  Or it might be that the fundamentals of your bu- business need to tweak, and that you need to focus right now more on your profitability and your margin before you keep on spending.
• 18:20
  That's a really tough conversation for someone to bring to the table when their job is to buy you media spend. That's what they've been hired for.
• 18:27
  That's where having somebody internal to look at the numbers and to truly understand them and what they mean is going to help you make that better decision and is going to help you understand not about what optimizations can make ads better- But about more holistically, is spending money in the first place even a good idea?
• 18:46
  Are we spending on the right platforms? Are we spending it towards the right e-commerce channels? All those larger questions need to happen inside a company. Um, they, they don't come from the agency.
• 18:58
  Even decisions like what channels to be spending ads on, do we go to TikTok? Do we go-- what, what about Snapchat? Mm-hmm. What about, you know, is Google Ads optimized?
• 19:08
  Are you suggesting that someone in your role can help advise the channel differentiation or spread, or isn't that the expertise of the paid acquisition, paid media agency? This is the new role of data analysts.
• 19:21
  We're spending less time heads down at a desk in Python, in SQL, building dashboards, building reports, answering ad hoc questions.
• 19:30
  All that extra time now is put into helping get to the answers and the business outcomes that can be driven by that data.
• 19:42
  So today it's so much faster for me to get to the numbers I need to help brands make decisions. My job is to be asking the questions that are important for making the right decision.
• 19:54
  It's about looking at deciding which KPI actually matters for this goal and connecting those two and making sure that KPI is the one we're tracking, because sometimes that KPI can be buried even if it's the most important KPI for what you're working on today.
• 20:09
  So the analyst's role have, has changed from crunching numbers to helping with the actual strategy and helping leadership teams understand what role data plays in shaping that strategy and shaping those future outcomes.
• 20:29
  If I take it from the perspective of the business owner, that's where a lot of my conversations are at that level. Yep. And the questions come down to, at this time of year, can I increase my spend and do so responsibly?
• 20:45
  And that's, that's like a big open question. Yeah. Now, we, we met in part because we were matched up through the, the Skew accelerator program for CPG brands.
• 20:57
  We were working together last year going about this, going into fourth quarter, and the big, simple but complicated question from the owner was, "How much can I increase my meta ad spend?"
• 21:12
  It was kind of a simple question, but it made a consequence because that money allocation was gonna go to Meta or somewhere else, right? And there was this question around how does,
• 21:25
  how does that person reach a level of confidence that, you know, the paid media agency is saying, "Yeah, assuming we're gonna increase spend," but there's this
• 21:37
  emotional sort of tension or turmoil inside, but like, is it really the best use? Does it make a difference the more incrementality we wanna be seeing in the outcome? How do you,
• 21:50
  how do you balance the owner's desire to see a growth
• 21:55
  that paid-- maybe if, if it's a big enough company, it's got their own internal team running paid media, maybe they also have an agency doing certain channels in special ways.
• 22:06
  Where do, where does-- maybe you could describe a story or an example where a business owner is able to come back and say, "Oh," like some sense of relief. Yes. "Okay. All right.
• 22:16
  I feel like, I feel good about what's on the- Yes... on the credit card." Yes. Um, and the example you started with is a great one that I'll, that I'll dig into.
• 22:24
  So in that case, their data was telling a story as well as-- and when they were, they were talking to Claude, it was confirming it, that once they spent more money, their ROAS was increasing.
• 22:36
  And leading into Black Friday, they had frozen their spending, and they said, "I'm nervous about spending more because my ROAS is going up, and everything is telling me about this, if I want to spend more, I'll have a higher ROAS, but how am I supposed to grow
• 22:52
  if every time I spend more money, my ROAS goes up?" So the first thing that I did was I looked into what is their profitability for ads today. And rather than asking, "Is their ROAS going up, and can we prevent it?"
• 23:05
  I said, "Is it okay if ROAS can go up? And how high can ROAS get while they still make money?"
• 23:11
  The metric that really mattered for that company at that time was the profitability per acquisition compared to the average cost of acquiring each customer.
• 23:23
  So essentially a new customer cost of acquisition, um, and then layer-- comparing that to their profitability. And what we found was they had rema- still making a decent amount of money per purchase.
• 23:35
  So a customer would come, it would cost some money in ad spend. However, they were still profitable. Their ROAS was going up, but it was just closing that gap a bit.
• 23:45
  So at this time of year when they can acquire new people, when this company has a phenomenal a, uh, retention rate and they would bring people in for a second purchase, that would all be profit, my advice was, "Let ROAS go up.
• 23:58
  Here's how high ROAS can get for you to break even on the first order.
• 24:02
  If you get up to there, you'll be making all of your money on the second, third, fourth order, which you know you will because customers always come back. They love this product."
• 24:11
  They were able to use that to confidently start spending more money, track their goals on the KPIs that mattered, which was this profitability versus the new customer acquisition cost, and they were able to acquire more customers during the holiday season than they would have otherwise had they just focused on the metric KPI they were looking at, which was ROAS, which was on an upward trend.
• 24:33
  So as a result, they made more sales, they have a larger customer base, and they made more money the following year in the spring as those customers came back for a second, third purchase.
• 24:43
  I think the big Question that was also behind the question was, at a strategy level, the business owner was asking, "How much effort do I da- give to going into a retail store environment?"
• 24:57
  And that was, I think, initially the biggest appeal. Like, "Oh, it sounds like a new channel, this new opportunity.
• 25:04
  There's prestige associated with getting placement in this store, and, like, maybe that's where I need to be focusing as I go into this peak period." But it was interesting to see how with the confidence that came of
• 25:18
  what was already working, what was already selling through Shopify, there was an ability to sort of, I would say, set aside some of the, um, potential distraction. You know, it's a critical time of year.
• 25:33
  We can only c- execute on so many different channels at one time.
• 25:37
  It was interesting to see the sort of relief and the ability to sort of settle in and follow through on what was right in front of the business owner as an opportunity and just needed more attention rather than spreading that attention across different ones.
• 25:52
  Mm-hmm. Does that make sense to you? Yeah, it does. And I think that really gets down to having that person in place who understands the data. It's someone who can be held responsible.
• 26:01
  It's someone when they say, "Let your spend go up to this number, and you'll be fine," they're responsible for it.
• 26:08
  They-- At that, at that point, I had put my, I had put my reputation on the line for her to say, "You can spend this money, and you'll be good to go."
• 26:16
  You-- It feels a lot different when that's coming from someone who you trust than when it's coming from between a, a, a chat. And when it comes from a chat, you say, "Are you sure?"
• 26:26
  And you're gonna continue [chuckles] checking it probably daily. Um, and unfortunately, the answer will change, just as the nature of, of that, uh, technology, um, the way it works.
• 26:35
  Like, it, it will shift, and we'll say, "Well, I'm not sure. Maybe." But you still have people who you can rely on and trust to say, "This is it. This is the math. You can stop worrying. Go focus on retail.
• 26:48
  That's the bigger opportunity. Here's one number you're going to look at every morning. You don't need to spend an hour digging into the numbers. You don't need a back and forth. You don't need to second-guess yourself.
• 26:57
  Watch this KPI. Here's where it can reach. As long as it's-- When it reaches that, just dial back spend and go ahead and spend your time on this bigger opportunity out there."
• 27:07
  What gives you the, the right to make those kind of big calls? Who are you to make that kind of judgment? Tell me more about how you got into this whole business and where it is that it finds you today. Yeah, absolutely.
• 27:19
  So I've been working, um, entirely with CPG brands now for six years, uh, focused on the Shopify, Amazon ecosystem, Google Ads, Meta Ads, um, from Fortune 100s down to, uh, mom-and-pop operations all across the board.
• 27:35
  Um, it's amazing to work with a huge variety of brands.
• 27:38
  Like I said before, it's less about the size of the brand for when they're looking to work with someone who can help understand data and more about the internal team and what it looks like.
• 27:48
  So, you know, you have a-- you have brands who might be with, um, a multinational and teams that can't get a data person, and, and that's where they can work with someone like me.
• 27:57
  You also have mom-and-pop shops who, um, might have come from the creative side, from the product invention side, and need someone to help with those numbers and e-commerce growth and understanding Shopify.
• 28:07
  Uh, before in my career, um, I-- So I came from an MBA background, uh, at Georgetown. I used that to really understand how you connect data with business problems.
• 28:19
  Uh, from there, worked in tech for a comp-- for companies doing the, a similar thing, understand your customer, understanding opportunities in the fitness and hospitality space.
• 28:27
  I had a short stint at Meta helping during the, um, post-election incidents in the mid-twenty tens, uh, before pivoting over to full-time working with brands, small and large, and doing what I'm really passionate about, which is understanding how data can actually help your business and can be a value add for your business.
• 28:47
  Um, so that's where at any time we're working with a handful of brands, helping them understand those numbers.
• 28:53
  Um, and it's, it's obviously changed over the years, but it's still just as exciting today as it ever has been. Yeah.
• 28:59
  Well, again, I'm looking forward to spending some time with you in Austin in the workshop that we have coming up.
• 29:04
  But for someone else who already has a team and they're trying to decide where they spot that person on their team who might-- they might give more responsibility for the data, they show signs of, of being able to, you know, deliver more in terms of the kind of analysis and, and process you're looking at here for yourself,
• 29:25
  what are the characteristics of someone who's good at this kind of work? Great question. Um, I would say it's, it's likely going to be someone who defines themselves as, as a data professional is what you're looking for.
• 29:38
  Uh, on teams, I'll usually have a point of contact who I'm working with to help them become more self-sufficient, and that's somebody who understands the numbers, likes to look at a dashboard but not, might not understand the nuts and bolts.
• 29:51
  If it can be done internally, typically, a team has a data team already, and it's gonna be a data analyst, a data strategist like myself.
• 30:01
  Um, there's a bunch of other data roles that won't be the topic here, but those, those people are starting to also spread more into the strategy as, as their roles get simplified, or not simplified, but, uh, changed with AI.
• 30:13
  I would say a lot of teams, though, you are looking for someone who can help with the nitty-gritty of the data. That's where someone like myself comes in.
• 30:22
  That's where looking for someone who is a data professional makes sense for a lot of companies, uh, even at the smallest level, even if it just means at least making sure that you have a reality check and that all your ducks are in a row, I think it can really help out.
• 30:35
  I appreciate the idea of people being already a data professional. But for someone who maybe they've got an existing team, and they're wanting to spot the potential to develop more ability in this area-
• 30:47
  Is it just a proficiency with numbers and say they're good with spreadsheets, and so that's the cl- that's the sign or signal? Or is there something else?
• 30:55
  Because what I heard you say earlier is that today with AI, there's less gathering and crunching of numbers. There's more analysis, and therefore a strategic thought process that needs to be there as well. Yeah.
• 31:06
  So it's almost like there's a, there's a bigger understanding someone needs to be effective. How would you speak to that, or how would someone spot that on the team they might already have? Yeah, that's a good question.
• 31:19
  I would say you're looking for someone to develop into a strategy role at that point because that is what the data value add is in 2026, is very much the person who's connecting the nuts and bolts of the numbers to the business strategy.
• 31:35
  You have to pull that person in as well, someone who you're willing to bring in to say, "Here's what the company is trying to do" and have those real conversations about what does the data say that we can do to reach that goal?
• 31:49
  How can we be using data to measure towards it? Um, this person is definitely comfortable with numbers. They're comfortable in Excel. Um, they're comfortable on your...
• 31:58
  because a lot is happening obviously in SaaS products as well.
• 32:01
  They're always getting into the numbers and the dashboards on the tools you already have, and helping leverage those people with just sharing more information about what that business strategy is and making sure that they're aware of where the business is going so they can find those opportunities to say, "Here's where data can help."
• 32:18
  Um, when a question comes up about strategy, they can say, "Have we thought about looking at the data in this way instead?
• 32:24
  We have this report that we've produced that really focuses on this specific area that might be around, you know, retention.
• 32:31
  But have we thought about how acquisition needs to happen first in order to build up a retention engine? Let's look at the numbers over there as well."
• 32:38
  That's the person you're looking for, someone who is both, uh, technical as well as strategic. Now, in Austin, I, I host the Shopify meetup, and people come through there all the time.
• 32:48
  The other day, someone, uh, asked, asked about this question, and they said, "Well, I've got Triple Whale. That's all I need." And maybe it's someone else would say, "I got Northbeam," you know, whatever.
• 33:00
  Uh, how do you respond to that? Is that sufficient to be making these kinds of decisions? What's your take? That's a great question. Um, so you named two companies that are great. I partner with both.
• 33:12
  I have a-- most of the brands I work with will have a tool like a Triple Whale, Northbeam, or others that they're, that they're using. Um, what those tools-- Those tools give you two things.
• 33:22
  Those two in particular do great things for, for attribution and improving how you're actually tracking, um, and, and giving attribution to what ad dollar is driving which sale, so there's value right there.
• 33:35
  And then you have the dashboarding and reporting side that both are quite strong in. Both are going to be very good at serving up the numbers and giving you all the numbers that you might need to make decisions.
• 33:47
  I've had people come back to me and say, "So what's the data analyst's role now that these tools can do that?
• 33:52
  And now they also have their, their AI bots, and I can just ask them questions, and they'll tell me whatever I need about my data. What happens to data analysts?"
• 34:00
  And my answer is that that piece of the job was never why you had data analysts in the first place. That was not the value add.
• 34:08
  That fell to the data analyst for certain, but before we had, before we had, um, AI to help people self-service better, but ultimately, that was not the value.
• 34:19
  The time when a data professional really earns their keep is when you hit the problem, and you say, "Which piece of data do I use to answer this, this question?
• 34:30
  Because I, I'm looking at my dashboard, I have these ten metrics. It's not clear because my CAP's going this way, ROAS is this way, AOV is this way, new customers are that way.
• 34:39
  It all looks somewhat similar to last year, but also we're not performing the same as last year. The AI is telling me this, but it just doesn't seem right because I already thought of that."
• 34:47
  That's where the data analyst today, and always in the past, has earned their keep because they come, and they say, "Here's how we're gonna cut it down. We have our numbers. Now let's give some context.
• 34:56
  Let's start segmenting the numbers differently." And thinking through how to segment, thinking through what context is necessary
• 35:05
  is the important piece, and that's what brings a team that has a Triple Whale or Northbeam that they're using, um, effectively to understand their numbers.
• 35:16
  A data specialist is what brings them to the next level of not just understanding the numbers but actually using their numbers to make data-driven decisions confidently and making better decisions through data rather than simply having the insights.
• 35:30
  There was one piece of the puzzle that we
• 35:34
  understand is important in terms of what's in Google Analytics, and you've described that there's, there's one sort of missing piece that you need to add to your stack to make Google Analytics perform optimally.
• 35:44
  What's that? Um, yeah. So Elevar now, uh, I'm sorry, Audiense, formerly Elevar, um, is one of the tools out there that I've partnered with quite a bit.
• 35:52
  Uh, there's some other great ones as well, and what they'll essentially do is make sure that you have a data layer under-- I'm sorry, a, um, yeah, data layer underneath that that's feeding in the correct numbers to close the gap on Google Analytics.
• 36:06
  Uh, tracking in Google Analytics is everything is different today than it was in the past, and you wanna make sure that you're feeding that data correctly so that you have a tool that's showing you accurate data.
• 36:16
  Um, Google Analytics can still be very useful even without it, but you do need somebody who understands
• 36:23
  why it's okay to only be-- for it only to be receiving eighty or eighty-five percent of your transactions, which is okay.
• 36:30
  We can still make really good decisions on a, on regular GA four, but it's pretty uncomfortable for an owner who says, "I hear you, Dan, but also twenty percent of my transactions are missing."
• 36:41
  Um, so there's tooling that helps you close those gaps, helps you send the data back and forth between these platforms, um, quite a bit better. So that's an important piece of it as well.
• 36:50
  For a company that is crossed the hundred million dollar mark- Mm-hmm... they've got their meta optimized in. They've done Google Analytics long, long ago.
• 36:59
  Triple Whale, North Beam, whatever the checklist is, that's already taken. But maybe at this point they are wanting to explore new ways of reaching customers, creative ways.
• 37:11
  Um, you know, the, I came across, uh, a conversation earlier this week with a brand that is using a sweepstakes and trying to explore how that can be used for customer acquisition. Sure.
• 37:22
  Um, there's all kinds of creative ways of trying to tap in and find some other new opportunity. When you hear
• 37:31
  that more unusual activity happening, does that change how you think about measuring and making sure everything is accounted for from an attribution standpoint?
• 37:42
  When we're thinking about a hundred million dollar brand, I think there's two pieces to that. Um, the first thing I think is they have so much data to work with, they can make such smart decisions with it.
• 37:55
  Um, by that point, I think every brand that's doing hundred million should have a data, data specialist in-house, a data analyst or data strategist, um, because at that point, you're driving purchases so fast that decisions that might take a small team months to build up enough data to make a decision on, a medium-sized team, you know, once you're doing one or two million, you can start making good decisions with a week or two of data.
• 38:19
  When you're doing a hundred million, you can be, you can be pivoting daily and making really good decisions, and that's specifically hundred million on a sp- on Shopify, let's say, on a channel.
• 38:30
  Um, a lot of the brands I work with are doing far more than a hundred million in retail, but their D2C is a much smaller slice, and that's, and then they look more on the D2C side, like a smaller brand.
• 38:41
  So what I would say then is they are able to test a lot more and a lot, and able to take more risks at a hundred million plus because you get the data response much more quickly and much more, um, at a much deeper level.
• 38:57
  For a small team who might have on D2C five million and they wanna do a sweepstakes, they wanna understand if it works,
• 39:04
  they're going to need to run it, let's say, for a month or so to build up enough data to really know, and then they need to separate that out from all the other things they did that month.
• 39:13
  So they released a new product, and they made these tweaks to the website. They had the sweepstake, and they're trying to peel it all apart to understand if the sweepstake worked. That's really tough to do.
• 39:22
  A hundred million dollar brand, though, can really run straight into some of these initiatives, test it for a few days, get a very clear lift, and slot it in between their other initiatives.
• 39:33
  So then it's really about speed and which way you have to make decisions. The more you sell, the more quickly you can pivot and adjust, right? Yeah.
• 39:41
  When you're making data-driven decisions, you're looking to get enough data points, which is essentially, you know, if you, if you wanna understand if conversion rate is going to increase, uh,.1% from something, you need to have enough conversions to be able to break that out from normal behavior.
• 39:58
  That's going to be a large number that might take it, for a small team, if you need thousands of conversions to make that decision, it might take you a week to get there.
• 40:09
  For a much larger D2C business who is doing a hundred million on D2C, they, they're getting a thousand purchases, you know, by, by the hour.
• 40:19
  So then it might just take a few days, and you have enough volume to, to get down and gritty, um, to a.1% conversion rate, to a small lift in AOV.
• 40:27
  You can see all that much more clearly, much more quickly because you're getting more data points that let you make those, um, kind of get visibility and granularity into it.
• 40:38
  I wonder then if I could take the, um, the issue of the brand under five million. Yes. For that situation, we're talking about paid acquisition and traffic. Let me see if I can share my screen. Sure.
• 40:53
  And let's talk about something that came up just as we were preparing to, to talk about this. Now, I've got a, a f- a friend in Austin, Hal Smith from H Street Digital, focused on paid acquisition.
• 41:07
  We complement each other at our eBusiness PROS agency focused on retention and helping those two things balance each other. So Hal and I go back and forth. He pulls out a comment from Cody Sanchez- Mm-hmm...
• 41:20
  also in Austin. And Cody says, "I'll die on this hill. There's absolutely no reason you should touch paid ads until you're doing five to 10 million in revenue."
• 41:31
  And she goes on and explains that, you know, a lot of marketers hate this advice because marketers make money when you do paid ads, and we talked about that a little bit earlier, who you, who you're gonna trust to grade the work.
• 41:43
  But Hal agrees, and the comments on here are interesting. I called you out, and tell me more about your answer here. Um- Yeah... how do you respond to this question, particularly for the brands under five million?
• 41:58
  What do you have to say about this? Absolutely. I, I think the, the comment will leave a lot of brands under five million scratching their heads saying, "Well, if I don't do paid ads,
• 42:07
  then how, how am I getting to five million? How am I actually doing it?" I know she provides some recommendations on organic, um, and there are some brands where that can work for.
• 42:16
  In, in the end, I think I would say it depends, like so many questions in this space, and there certainly are some categories- Come on, do, do, do better than it depends. Come on, Dan.
• 42:24
  Well, I'm, I'm, I'm gonna, I'm gonna go. I'm, I'm d- defining it.
• 42:27
  So if you're, if you're in a category where people already know the product, they know the purpose, they're very-- they, they know they wanna buy it, then you don't need paid ads.
• 42:36
  If you already have customers who can flock, you just need to get your product in front of them, then you're good to go.
• 42:41
  But some of the most innovative small companies I've come across, people aren't, people aren't there yet. They're not aware of this category.
• 42:49
  We've seen some in SKU where people would never think they need this, and they discover it, or they taste it, and they say, "Oh, wow, this is amazing."
• 42:57
  How else paid ads can still play a really important role potentially. What I would say is you have to be careful about your metrics, like I said before. ROAS, really dangerous metric to use.
• 43:08
  Um, it doesn't necessarily tell you what you want. It oftentimes is going to blend in returning customers and look-- may, you know, you hear all these ROAS performance metrics.
• 43:17
  You need to understand it if you're gonna use it.
• 43:20
  Where brands, I think-- where I think they were-- that conversation was really getting, getting r-right and where, or I'm sorry, where I think it was starting to get was that brands can spend money, but they need to be doing so responsibly, and they need to know what's going on.
• 43:35
  When the thought about, you know, it's hard to get a paid agency to do good for you under five million is that it's really easy to start giving money to a paid agency
• 43:45
  and-- or even just to do it yourself, to start paying money and to keep on feeling like, "Okay, let me just keep on spending to see if I can make it work. Let me optimize it, let me tweak it.
• 43:53
  I'll make these small changes and see if it works." Well, if you are losing two hundred dollars for each acquisition on paid ad media spend, and you make forty dollars per customer,
• 44:05
  you can make all the small tweaks that you want. You're not gonna become profitable. You're not gonna start making money. You need something massive to change.
• 44:12
  And I think the fear is a lot of these companies under five million are spending money, are losing hundreds of dollars per acquisition, and are thinking about solutions that might mean they lose a few dollars less than hundreds of dollars.
• 44:25
  But they're not gonna become profitable.
• 44:27
  So you need to think-- you need someone in-house who is giving it that scrutinization, and you yourself as a founder need to be ready to make a difficult decision to potentially turn off ad spend if it's not working.
• 44:38
  Don't just keep on running forward. Don't keep on spending money. Um, under five million, you're still asking questions like, "Is D2C the right channel?"
• 44:47
  That should still be on the table until you hit five million, and not every company can hit five million on D2C. A lot of them can go far beyond that.
• 44:57
  Some of them, it might be about hitting one, two million, getting-- proving that there's demand and then going to retail. For others, there might be a good opportunity for Amazon.
• 45:07
  Before you're at five million, you still have that tension, and you need to be able to face it, and you need to know what questions to be asking to really grapple with that decision of
• 45:18
  is ad spend-- is there a future in ad spend? Is it how I'm using the money, or is it just that the money's, money's not gonna get me there?
• 45:25
  Well, I think these are the bigger big questions that is much more than just about dialing in a ROAS number or making a few savings or efficiencies.
• 45:34
  I mean, I was at a, um, event recently where a certain maker of a certain bag of chips fried in beef tallow was describing how D2C and Shopify and their ad spend helped get them to a certain viability.
• 45:50
  They knew that people would buy it at a premium price. Yes. But now, bag of chips is big and bulky and, you know, it's not heavy, but boy, it takes up a lot of space. Hard, expensive to ship.
• 46:04
  Now they can move into the retail channel because the mar-market proved people would buy it, and so their priority going into this fourth quarter is all about the retail channel. Yes.
• 46:14
  And so I guess that's a nice example of how e-commerce is not an end in itself, right?
• 46:18
  We do it for a good reason, and then we can make new decisions as the data informs it, but also as we start to understand the bigger opportunities, they might be in other channels. Yes.
• 46:29
  With brands under five million, typically I'll first sit down with them and just have a point-blank discussion to understand if D2C is a growth channel for them.
• 46:38
  And when I say growth, I mean that you can have long-term success and continue growing over time. Um, and with small brands, there are some that I come across where I say, "Wow, you can just be printing money.
• 46:48
  You, you have something that works here." Um, and that usually means you have a high enough AOV to support a quite expensive market for buying advertising.
• 46:57
  It might mean that you have such a high retention rate and such guaranteed repeat purchase that even if you lose money on the first order, you'll make it later. Don't worry about it. Spend a little bit more money.
• 47:08
  Once you get somebody in, you'll keep on bringing them back and making more money. Um,
• 47:14
  on the other side of it, you have some brands where D2C is a great early-stage channel, where D2C can help create the stories that they can then bring to retailers to get in the door. They can prove the market demand.
• 47:28
  That's what most of my time is spent doing. Even though I call myself a data strategist, most of it is spent having these conversations with founders and leaders to talk about where can we actually bring the business.
• 47:40
  And people tend to say, like, "Dan, you have such a unique perspective. You bring such a different view to the business. Why is that?" I say, "I, I'm, I'm just bringing the view of the numbers.
• 47:49
  I'm speaking for the numbers and the data over here and telling you based on just the data alone, what is that perspective that's coming to the table?"
• 47:58
  Um, and that's where you can really start to make some of those hard decisions about channel strategy, about how much to spend. In the beef tallow example, where should we, where should we pull it?
• 48:08
  How far should we let this go before we say, "Great, we have what we need. Time to go to retail"?
• 48:13
  Um, that, that's ultimately the value that somebody who understands the numbers and can truly speak for them brings to a business. Yeah.
• 48:21
  Well, as we wrap up here, you've had several years of observing the trajectory of how the profitability of these, you know, e-commerce channel can be for brands.
• 48:33
  What is your take on where we are right now going to the end of twenty twenty-six?
• 48:39
  How do you see the larger trend or climate, if you would have any perspective to share on what a brand may be challenged right now is trying to deal with? Has it always been this way?
• 48:50
  What are you seeing, and maybe do you have any forecasts for twenty twenty-seven? Yeah, for sure. Um, Black Friday used to be much more predictable. Now every year we're going into it- Feeling a little bit more unease.
• 49:01
  And where I think that comes from is changing consumer sentiment, and it's really hard to know where the consumer sits.
• 49:08
  As far as, you know, with econ- economic shifts, consumers are sometimes holding back their spending, sometimes leaning into it. Um, and getting a really good grip of...
• 49:18
  And whoever has the crystal ball of how will consumers spend this Black Friday is going to be answered that question better than anybody else. So I'd be ready for it to go either direction. Be ready for it.
• 49:31
  You know, if it- things are going well, what are you going to do? If things are going poorly, what are you going to do? And where data comes into that is be tracking your numbers closely.
• 49:42
  Watch for those inflection points where Black Friday picks up. And also, I think we're in, still in a period where I would be responsible about profitability.
• 49:50
  In the past, we had periods of e-com, just grow, grow, grow, spend that money, bring those people in the door, in the door. We're still today in a spot where profitability is being thought about a whole lot more.
• 50:02
  It probably shouldn't have been in the past, but now it is.
• 50:04
  Make sure you know exactly where those cutoff points are, where the business swaps from making money to losing money when you get to the bottom line, not just the top line. That's super important.
• 50:14
  In twenty-twenty seven, we're seeing improvements in meta right now more broadly. I think that we're also seeing TikTok start to work for more brands. Um, I...
• 50:26
  What we've seen the past few years is a major shift in how we approach marketing as the, as the things are becoming more algorithmic now that the algorithms are so much better.
• 50:36
  Uh, AI is, is allowing algorithms to be much more specific, so expect these pieces to start snapping into place.
• 50:46
  It's gonna be less picking and choosing where your money gets spent, and more handing your money over to these companies and saying, "I, I hope you spend this well for me." Um, that's gonna be great for some companies.
• 50:59
  It'll be disastrous for others. And I think it's gonna be a, a period of twenty-twenty seven figuring out which side you land on. As, do the algorithms like me?
• 51:06
  Are they going to do well for me, or are they just not getting it? Are they not putting the ads where they need to be?
• 51:12
  And as I lose options to do things manually, do I need to be thinking about a different way to drive my business? The sooner you're seeing those signals and making decisions, um, the better off the company's gonna be.
• 51:24
  Well, what I'm hearing from you is that it really comes down to measuring profitability at a level of detail that's appropriate to your sales volume, so that the speed at which you can take in new information, pivot, adjust, run tests or experiments, is appropriate to the scale of your business for the size that it is.
• 51:43
  And then I think I'm also taking away from this that you still need good humans somewhere to have these conversations around what is an evolving data story, and that's not something that a tool can just take care of for you.
• 51:56
  So Dan, I'm really glad to have you in the kind of work that I'm doing with brands. It always gives me more confidence when we're trying to make decisions.
• 52:03
  So thank you so much for talking about how you do this work, and come back again someday at the Frankly Ecommerce Show. Yeah. This has been great. Thank you so much.
• 52:11
  We hope you enjoyed this episode of Frankly Ecommerce with host Patrick Scott Pittman. Read episode notes and get links to learn more at franklyecommerce.com.
• 52:21
  Next time you're in Austin, check out the Austin Shopify Meetup. It's open to all.
• 52:28
  There's a good chance you'll meet just the right person, as it's one of the longest running meetups with twenty-six hundred participants going on fifteen years.
• 52:38
  If you'd like to get more involved beyond the meetup, then Frankly Ecommerce pass holders participate in workshops and special invitation-only events. Subscribe to our newsletter at franklyecommerce.com to learn more.
• 52:52
  This episode is sponsored in part by the friendly team at eBusiness PROS, a marketing firm in Austin, Texas. If you're a brand seeking to connect with an enthusiast audience and hold their attention,
• 53:05
  if you want less discounting and more story, if you're curious about what we might mean by attuned marketing as contrast to behavioral trickery, then contact us at ebusinesspros.com. Ask for Abigail.
• 53:18
  That's me.
• 53:38
  [upbeat music]