Latest / Shiny New Object - a Marketing Podcast / Episode 280 / Tom Boxall / Naked Wines / Performance Marketing Manager
Transcript
- 0:00Tom Boxall: Truly understanding what's moving the needle and
- 0:04where you should take that next is the next big question.
- 0:06Speaker: Are you a fitness loving foodie? Then defy snacks
- 0:15is for you. Defy offers real high quality chocolate you love,
- 0:19combined with functional ingredients, giving you as much
- 0:21as 26 grams of protein per bag. So you can be both a little bit
- 0:25bad and a little bit good at the same time as a female owned
- 0:28brand, defy dedicates a portion of its profits to other women
- 0:32owned businesses currently sold in the US. You can find us on
- 0:35defy with an eye snacks.com and follow us on social at defy
- 0:39snacks.
- 0:46Tom Ollerton: Hello and welcome to the shiny new object podcast.
- 0:49My name is Tom Ollerton. I'm the founder of automated creative,
- 0:52the creative effectiveness ad tech platform. And this is a
- 0:55weekly podcast where I get to interview one of our industry's
- 1:00leaders. And it's an absolute pleasure and a privilege to be
- 1:03able to do that and learn each week. And this week is
- 1:05absolutely no different. I'm on a call with Tom Boxall, who is
- 1:10performance marketing manager at Naked Wines. So Tom, for anyone
- 1:14who doesn't know who you are and what you do, could you give the
- 1:16audience a bit of background?
- 1:18Tom Boxall: Of course, firstly, thank you so much for having me,
- 1:22absolute pleasure. I've always surrounded myself with food,
- 1:28wine. I think it's one of the greatest things you can do in
- 1:32life. I truly believe that all good things happen around a
- 1:37dinner table, and so that's where I've kind of better fitted
- 1:40myself in, into an industry. It's one I truly believe in. But
- 1:44professionally, I like to always say I've kind of trod the path
- 1:48between data and design. So whether that's from a young age,
- 1:52studying both art and maths or English and business at uni,
- 1:57I've always loved finding out the numbers behind why a piece
- 2:00of creative or branding has worked, so letting the data
- 2:04paint the picture, so to speak. And I guess I realized these
- 2:08intricacies coming into marketing through CRM. At my
- 2:13first role was even younger, young CRM exec, working hands on
- 2:19with data, customer segmentation, building out
- 2:22flows, intricate ways of measuring performance, but
- 2:25ultimately designing creative and messaging across multiple
- 2:28channels that speaks to the right audience at the right
- 2:32time, and from there, this naturally led me into growth
- 2:36marketing. So I've since worked with small startups on building
- 2:41their growth strategy where the MO is, grow, grow, grow, what,
- 2:45no matter what the cost to now working at Naked Wines and tying
- 2:50together our creative strategy and performance marketing.
- 2:54Tom Ollerton: Tom, no one has started this podcast by saying,
- 2:57I like to put myself in the center of food and wine and what
- 3:04an intro you sound like an automated creative person.
- 3:07That's quite funny. So let's get into it. So in that career, what
- 3:11has been the best investment of your own money, your own time,
- 3:15or your own energy?
- 3:17Tom Boxall: I think, put simply, is to understand the customer.
- 3:20Now that can be a very easy thing to say. It's quite a
- 3:24common thing to say, but most businesses have only scratched
- 3:29the surface. Most companies know what the product or service they
- 3:33are selling is. They know what the benefits are. They know how
- 3:38they intend their ideal customer to use it. But not many truly
- 3:43know why the customer bought, and particularly in high growth
- 3:47startups or any business for it's maybe high performing
- 3:51growth function and fast moving, it becomes very, very easy to
- 3:56chase growth at any cost, or, you know, Chase growth because
- 4:00the row assets are good in that channel. The trouble is, your
- 4:05customers change your target demo changes. Your product may
- 4:09change, you know, bottom of funnel, creative may become more
- 4:13expensive, more competitive, and do you therefore still know who
- 4:18that customer is. I think, in the world of performance
- 4:22marketing in particular, it's moved away from the bespoke
- 4:26targeting and intricate campaign structures that there once were.
- 4:31You know, you can no longer contact or you can no longer
- 4:35target, 40 year olds called Diane who live in a cul de sac
- 4:40in North London, who have a dog, or in middle, you know, middle
- 4:44income, that world is completely gone, that will continue to
- 4:49trend in that direction. And that's where I believe, and
- 4:53where my world kind of collides with it is where creative
- 4:57becomes new targeting. Now creative as for new targeting.
- 5:01That's a phrase which you've probably heard everywhere. You
- 5:05may have even heard your meta rep spout it, and I apologize to
- 5:10my meta rep if he's listening, but I tend not to believe any
- 5:14bullshit they say. But this is one line which I do truly
- 5:19believe, and I've seen work. You know, if you are taking any
- 5:23product or any anything that you sell, people will be buying it
- 5:28for different reasons. You know, let using wine as an example for
- 5:32Naked Wines. We know that you know yourself. Tom, you might
- 5:37buy wine for one reason, which is different to the way I buy
- 5:41wine. Someone might buy for the, you know, fancy label, because
- 5:46it's got a chateau on the front. Another might buy, you know,
- 5:50only Malbec, because that's all they ever drink and that's all
- 5:55they ever trust. You know, it's comfortable for them. And others
- 6:00might like to try something new. So how are you appealing to all
- 6:04of these customers? What is that buying behavior that your
- 6:09customers do? What is the funnel and journey that they go on to
- 6:13decide to purchase?
- 6:21Tom Ollerton: So the best investment of your time has been
- 6:23to really understand this customer. So how do you do that?
- 6:26What's your process for making sure that you are understanding
- 6:31them and marketing to people who have completely different uses
- 6:33or focuses on buying something like wine?
- 6:36Tom Boxall: Yeah, there's a few, few ways, from the ground up you
- 6:41can be building out post purchase surveys into your
- 6:44product, into your customer flow, understanding your Trust
- 6:49pilot pulling that, extracting all the data, and very easy ways
- 6:53to pull out your trust, pilot reviews and sort them by the
- 6:59word count. That's kind of an interesting way of looking at
- 7:02it. But if you sort by the highest word count, you often
- 7:05find the largest story, the biggest piece of information or
- 7:10nugget to unpack, and they really start to go into that
- 7:13story of, why did I buy I bought this because X. And then there's
- 7:18kind of deeper customer level research as well, so
- 7:22understanding what your core demographic is, your key
- 7:24profiles, and going back to my early days of CRM, we were
- 7:29actually building out these key profiles and customer
- 7:32segmentation and appending those on to each and every customer
- 7:36ID, so that from there, we could identify The types of products
- 7:41or venues or where they shop, what they interact with, what
- 7:44are their motivations, their likes, dislikes, behaviors, and
- 7:49from there, you then build in, okay, well, how does our product
- 7:52solve a solution for them, and that will give you a better
- 7:55understanding of how to target them, how to speak in their
- 7:59language, where you even are communicating what channel and
- 8:03what interactions you have.
- 8:04Tom Ollerton: So if I was going to push you on one singular bit
- 8:08of advice that you would give other data driven marketers to
- 8:11be better at it, what would that be?
- 8:13Tom Boxall: Think one of the biggest learnings I've had kind
- 8:16of growing up in this industry, growing up around data, is
- 8:22simply to back yourself and then go and test so you know, there's
- 8:29so many ideas and new theories running around, and you will
- 8:33only ever know what works for your business if you go out and
- 8:36test it. But in order for you to go and test that, you need to
- 8:40pitch the idea and get buy in, often from from key
- 8:45stakeholders, and that can be quite a kind of looming task to
- 8:51do. So, you need to find a way to tell the story within the
- 8:57data. I did a course years ago with Salford University called
- 9:05Beyond data analytics, and it was great fun, two, three day
- 9:09course, part of the time was kind of mucking about playing
- 9:13with blue tack and props and different charts to try tell a
- 9:18story from just one sheet of data. And what that gave me is a
- 9:23real perspective of, how can I pitch this new found, let's call
- 9:29it a shiny object. How can I pitch this to the CEO, the CFO,
- 9:34the board, or, you know, whoever you're trying to impress. What
- 9:40data or what goal will they respond to? And how can I align
- 9:44this task or this gut instinct to that? So if their goal is
- 9:50bottom line revenue, how will the test that I'm working on get
- 9:54them there? And that simply is the best way to pitch an idea.
- 9:59And. A lot of that comes from being personable, or, you know,
- 10:04digging into what they want to see, asking the right questions,
- 10:07getting to know their gripes. And that often comes with
- 10:11confidence, but if you are able to back yourself and ask those
- 10:14right questions, and, you know, engage with someone on a human
- 10:18level, find out what what's their difficulty, what's their
- 10:21barrier? Right now, you're halfway there. Once you pitch
- 10:26the idea, simply go and test it. And as we know, with testing,
- 10:31the winner may not always be the one that you expected, but
- 10:34you've made a step forward, you've learned, and you can go
- 10:38from there.
- 10:45Tom Ollerton: This episode of the shiny new object podcast is
- 10:48brought to you in partnership with madfest. Whether it's live
- 10:51in London or streamed online to the global marketing community,
- 10:55you can always expect a distinctive and daring blend of
- 10:58fast paced content, startup innovation pitches and
- 11:01unconventional entertainment from madfest events, you'll find
- 11:04me causing trouble on stage, recording live versions of this
- 11:07podcast and sharing a beer with the nicest and most influential
- 11:10people in marketing. Check it out at www dot madfest
- 11:14london.com.
- 11:21So your shiny new object is incrementality. So not everyone
- 11:27listening to this podcast can understand what you and I might
- 11:31mean by that. So why is incrementality your shiny new
- 11:35object? And give us a quick introduction to the topic, of
- 11:38course.
- 11:39Tom Boxall: So it's quite a recent one. It's almost like a
- 11:44hot topic in particularly growth marketing right now.
- 11:49Essentially, it's finding out which channel is working the
- 11:52hardest, where should we be investing our media spend, and
- 11:57what is driving the most conversions that wouldn't have
- 12:01happened anyway. Simply put, you know a channel could be
- 12:05performing well, you could be seeing incredibly high ROAs. But
- 12:10how do you know that this is not chasing people who were about to
- 12:13buy anyway, or even existing repeat customers, rather than
- 12:17finding net new incremental audiences? And what
- 12:22incrementality does is it allows us to truly know where we should
- 12:26be investing, where we can grow, and where we are wasting
- 12:29dollars. And so the next question will clearly be, how do
- 12:34you find this? Now the answer can be found through geo lift
- 12:39tests, mixed media modeling and multi touch attribution tools.
- 12:45If any of the big tools and providers are listening, then
- 12:49feel free to sponsor me. But I'll give you an I'll give you
- 12:54an example of how this has impacted us. So for the last two
- 12:58years at Naked Wines, we've been working tirelessly to build paid
- 13:05social from one of the least invested channels in the
- 13:08business to one of the largest paid channels. And that is no
- 13:13mean feat, and that comes through rigorous creative
- 13:17testing, creative velocity, creative diversity for anyone
- 13:22out there in paid social you know, my hat goes off to you,
- 13:26and it sounds daunting, but you probably need to be testing and
- 13:30launching way, way more than you ever thought. You know, 30, 40,
- 13:3450 creatives a week. The more shots at goal you have, the more
- 13:37learnings you get, and the greater chance of finding a new
- 13:39champion creative you'll have. And if that's really daunting,
- 13:45I've got quite a good way of understanding what is
- 13:49incremental and what is quite a good test for you to do in your
- 13:53paid social ad accounts. So I encourage you to pull your meta
- 14:00spend by individual creatives for the last 90 180 you know,
- 14:06days a year. And if you have iterations on the same creative
- 14:10group, these by their naming convention too, and from there,
- 14:14build out a cumulative distribution function to
- 14:17understand what total percentage share of ad spend was behind
- 14:20each and every ad, and you will be shocked to see how the
- 14:25performance and scalability of your ad account hinges on just a
- 14:29few champion creatives. So for example, in one market, in the
- 14:34last six months, we launched over 200 ads, just 10 of those
- 14:42were responsible for 80% of the spend. Just five of those ads
- 14:48counted for over 50% and just three of those ads counted for
- 14:53nearly 40% so that's pretty scary stuff, and that is how
- 14:59kind of pivot. For your paid social accounts can be and how
- 15:05much of it hinges on good creative performance. Now I tell
- 15:10that story because that is how we got to scaling meta to where
- 15:16it is now. But the trouble with scaling paid social to be one of
- 15:20the highest performing and most invested channels is we were
- 15:24working with assumed numbers. You know, we were assuming the
- 15:29quality of customers coming in based on a benchmark. Perhaps we
- 15:33were using, and you might do this too. Perhaps you're using
- 15:37predictive LTVs. You know, AI has got so good but we can't yet
- 15:42see five years into the future. So we're using predictive LTV
- 15:47models to work out our forecasting. And, you know, go
- 15:50from that stage top down to what does the next year look like?
- 15:54Maybe ROASes are even difficult to see. Perhaps you're a
- 15:57subscription brand and you know, ROAs, or you know, what it says
- 16:01in meta doesn't always add up to your to your bottom line. Or,
- 16:05you know, perhaps your How does your paid social channel sit
- 16:09within a full funnel mix? So what we did to identify the
- 16:13incrementality of our channels is run a geo lift test, which
- 16:18excludes a percentage of the market based on their buying
- 16:22behaviors, based on their kind of type of demo that they are.
- 16:27And the results that came back were truly staggering. And the
- 16:31truth was that it meant that our CAC on paid social was actually
- 16:37a third of what we originally thought it was. So
- 16:41incrementally, I now know that what the channels are that are
- 16:47working for us, what the channels are that aren't working
- 16:51for us, and that enables us to scale at a much, much faster
- 16:55pace than we ever thought was possible.
- 16:58Tom Ollerton: So there is a lot there. Thank you for sharing the
- 17:02very specific examples and the numbers as much as you could. So
- 17:05to play that back to you, you're taking a channel and then you're
- 17:08doing a showing one audience ad and a another Geo. You're not
- 17:13showing them the ads, and then you're working out the the
- 17:15relative difference between the two in terms of performance. I
- 17:18mean, that sounds incredibly straightforward, so what? What
- 17:22makes this complicated? Why can't Why isn't everyone doing
- 17:24that?
- 17:24Tom Boxall: Well, it's, yeah, it's a it's a very good point.
- 17:27You know you need, you need some tools to help do it, or you may
- 17:32need some, some geeks, to work out the numbers of who's who and
- 17:37what does a market look like, or where matches up where. But in
- 17:41principle, yes, you are excluding one part of of your
- 17:45target market, and you are understanding whether what's the
- 17:50difference in their behavior now that your channel has been
- 17:54excluded from them. And this can happen on any channel, paid
- 17:58social is one example. You know, I would implore you to run the
- 18:03test across all channels where possible. So you know, one area,
- 18:08one select group of counties, may not be receiving, you know,
- 18:12parcel or DM, or you know, you might put the exclusions at
- 18:16account level on your paid social but essentially, you then
- 18:21understand, and you monitor what is the behavior of that audience
- 18:26that have been excluded from your marketing versus a baseline
- 18:30of everyone else. And if you start to see that actually that
- 18:35audience, you know we were previously shifting X million
- 18:39dollars into into this region, on on X channel, and that
- 18:43audience is still buying from us, then that channel isn't
- 18:47incremental. That channel isn't doing anything, those people
- 18:51would have bought anyway, regardless if you spent the
- 18:55money on that, on that region or not, and in other places, you
- 19:00may then find that, you know, actually this channel is really
- 19:04moving the needle. It's really getting people to buy. And it's
- 19:09not just targeting existing customers, repeat customers,
- 19:13people who would have bought anyway.
- 19:22Tom Ollerton: So now let's talk about MMM, MTA. Like, are you
- 19:29less favorable of those because they're a bit slower? Or help me
- 19:34understand how, how you balance that the geo exclusion approach
- 19:38with those two disciplines?
- 19:40Tom Boxall: I'm a big fan of both. I'm a big fan of all
- 19:44three, and I think they all need to be working in tandem with
- 19:48another. They each paint part of the picture. So for example, MTA
- 19:56will start telling you, you know, where the customers saw
- 20:00your saw your ads or your marketing material. How did they
- 20:05engage with them? It might tell you that where meta says this
- 20:10person clicked within seven days, actually, MTA will start
- 20:14to tell you, okay, they may have seen this ad, but they didn't do
- 20:18anything. They then saw another ad on their desktop. Maybe they
- 20:23went on YouTube and they watched, you know, an ad on
- 20:27there. Seven days have passed, so you're now outside of Meta's
- 20:31attribution window. Perhaps they Googled you and browsed your
- 20:36website, but then didn't make a purchase. And then, you know,
- 20:39maybe they then, after all that, received a flyer in through
- 20:43their letterbox and used the voucher code on there, and then
- 20:48in the back end, your data is saying, well, that direct mail
- 20:52with the voucher code that they got through the post should get
- 20:56all of the credit. And that is just simply not true, because
- 20:59you've done, you know, you've invested in paid social,
- 21:03YouTube, maybe connected TV, all these other channels which now
- 21:07aren't getting any credit. And therefore, you know, come around
- 21:10the next financial year and you're planning your budgets,
- 21:13you're going to be investing in the wrong places when, actual
- 21:17fact, what really, truly motivated them were the several
- 21:20ads before. So I'm a big believer in MTA, and then then
- 21:25you start to paint that picture and projection looking forward.
- 21:29So because I invested in these channels, what might that
- 21:33therefore mean based on machine learning, based on AI, based on
- 21:38previous customer behaviors, what is this going to tell me,
- 21:43and how can I predict what is the best mix of investment
- 21:47across multiple channels? Now, geo lift tests should just be
- 21:53running underneath all of this. Geo lift tests will give you
- 21:56those hard baseline facts of this is causing X incremental
- 22:02conversions that wouldn't have happened anyway in this
- 22:06particular market. And they can happen on a regular cadence.
- 22:11They can happen, you know, across different channels, at
- 22:13different times. And what you're doing essentially stress testing
- 22:17your theories on who is interacting with what and what's
- 22:22really working.
- 22:23Tom Ollerton: So, so what's missing for you right now? So a
- 22:27really succinct and lovely breakdown of the benefits and
- 22:32the how tos of three different ways of making sure that your
- 22:35money's going the right way. Like, what is the missing bit
- 22:38for you? Where are you like, Ah, if only I knew this?
- 22:41Tom Boxall: It's good question. I think laying out the benefits
- 22:46of each you start to see how they can interact, but often
- 22:51they can give you more questions than answers, and to truly
- 22:56understand where that you should be, then forecasting and how and
- 23:00Tom, unfortunately, we've got to leave it there with a profound
- 23:02when and when will, you know, intent ramp up. When will orders
- 23:08pick up? What will CPMs look like? You know, we're talking
- 23:13now in the middle of US elections and all sorts. Black
- 23:18Friday, peak, Cyber Monday, etc, Christmas. You know that
- 23:23seasonality is always going to be a big question when you are
- 23:29doing your forecasting, and these tools will only get you so
- 23:34far with that. What they can give you is, how much, how much
- 23:40do I trust that this channel is performing to its best ability,
- 23:46and therefore give you confidence in where you should
- 23:51be investing next? But joining those dots, that's always going
- 23:56to be down to us as digital marketers, as growth marketers,
- 24:02so truly understanding what's moving the needle and where you
- 24:07should take that next is the next big question.
- 24:16question. So if anyone wants to get in touch with you about
- 24:19anything we've talked about today, where would be a good
- 24:22place to do that, and what makes a message that you'll actually
- 24:25reply to?
- 24:26The best place will be LinkedIn, yeah. LinkedIn. Tom Boxall, I
- 24:31mean, my what we've just talked about for half an hour, I guess,
- 24:36is what would make a good message? You know, truly
- 24:40understanding your customer, right? So if you know, I will
- 24:44always, always do my best to respond. And if the message is
- 24:49particularly targeted to me and of interest in speaking my
- 24:54language, then, then that's the that's the best route for that.
- 24:59But, yeah. Thank you.
- 25:00Tom Ollerton: Brilliant Tom, thanks for your time.
- 25:08Hi, just before you go, I'd really appreciate it if you
- 25:13could take the time to write a review of the shiny new object
- 25:16podcast on Apple podcasts, or iTunes, whatever it's called
- 25:19these days, or whichever podcast provider you use, we're in any
- 25:24podcast, so it would go a long way for us if you could just
- 25:27share the Word and give us a bit of a support on those channels.
- 25:30That'd just be fantastic. If you haven't got time. That's also
- 25:32cool. And yeah, if you could tell your colleagues about the
- 25:35podcast, and also, if possible, don't forget to subscribe. And
- 25:39I'd love to hear your feedback, if you'd like to speak on the
- 25:42podcast or be a guest, or you think I'm asking the wrong
- 25:45questions, anything. I'd be super interested to hear what
- 25:48you think. So please email me. Tom at, automated creative.net
- 25:54that's Tom at, I'm not going to bother spelling it anyway.
- 25:58You'll work it out. Thanks so much.