00;00;22;04 - 00;00;54;01
Speaker 1
Hello and welcome back to this week's episode of Dig and we are joined today by Hilary Born Doll. She's the founder and CEO of Mixed Analytics. And she is going to drop a lot of knowledge today on what it means to maximize your marketing hour, why and how you can leverage market mix modeling. Honestly, I did a quick pre-interview with Hilary and I was kind of blown away just by the sheer volume of knowledge that you have in the space.

00;00;54;01 - 00;01;02;19
Speaker 1
So I'm excited to be able to use it for myself as the head of marketing here. But thank you so much for joining us today.

00;01;03;21 - 00;01;19;24
Speaker 2
My pleasure. It's great to be doing this chat with you on one of my favorite topics marketing, mixed modeling and also full circle. Because I did previously work with the founders of Digg and I have a lot of passion for the brand.

00;01;20;03 - 00;01;46;21
Speaker 1
So it's oh, thank you. Yeah. And I cornered Hilary basically at a conference and was like, Hey, hey, you. You'd be perfect on the podcast. So thank you so much for, for letting me wrangle you into doing this with me before we kind of dove in to mix and and sort of your approach to marketing mix modeling. Can you tell us a little bit about your background in marketing and insights?

00;01;46;21 - 00;01;48;23
Speaker 1
I know you come from places like Nielsen and Kantar.

00;01;50;06 - 00;02;16;17
Speaker 2
Yeah, I think for me it actually started when I was really young that I had a love of math and I also had a passion for art. So when I stumbled upon marketing as a field, that would include both, but then as a more focused area of marketing, going into marketing research, I was just like drawn to it right away.

00;02;16;24 - 00;02;48;07
Speaker 2
So I did a co-op program. I graduated from Leng School of Business at the University of Guelph, and part of the program was working at Merritt's time, the Lightstone and marketing research. And I realized, Hey, this is what I want to do for my career. And, and then actually when, when I worked after school, I worked at Merritt's and then Hot Sparks and eventually found myself working at this company called the Modeling Group down in Connecticut.

00;02;48;07 - 00;03;18;28
Speaker 2
And we were running all of the advanced analytics for Proctor and Gamble globally and I literally see this all the time and it sounds so cheesy, but I fell in love with the technique of marketing, mixed modeling, because I was infusing all of this math and data science along with ad ads and creative evaluation into one solution for clients to ultimately help them understand their marketing budget and make smart decisions going forward.

00;03;19;09 - 00;03;37;27
Speaker 2
And I've remained in this field. I always will be in this field because if you find something that you love to do and it never feels like you're working, I just always enjoy every single day studying these brands and advertising. So I feel like it's a lifelong journey for me that's very.

00;03;38;10 - 00;04;06;23
Speaker 1
Very different from my views on math. So I'm happy we have people like you in the world who are fascinated by how you can sort of leverage that to make it easier to understand marketing ROI. I mean, you sort of touched on what mom, our marketing mix modeling is, but would you mind just giving us a quick rundown?

00;04;07;20 - 00;04;13;11
Speaker 1
You know, imagine were ten years old near explaining it to to a ten year old. What does it actually mean?

00;04;14;02 - 00;04;19;20
Speaker 2
Yes. So a dynamic model like a Bayesian linear regression model.

00;04;20;26 - 00;04;21;18
Speaker 1
Where words.

00;04;21;19 - 00;04;45;10
Speaker 2
Y equals B? No, I'm just kidding. I'm not going to show it. So if I'm presenting to my kids and explaining to them, then I say, you know, when you're watching YouTube or you are snatching my phone to go on TikTok and you see these ads for these brands, what mommy's basically doing is helping to analyze, Oh, was that ad effective for a brand?

00;04;45;20 - 00;05;07;16
Speaker 2
Did it either help increase brand awareness or consideration? So a brand that maybe you want to have or, you know, close you want to buy for yourself or where or did it help actually drive a sale? Like when we're in the grocery store and you see that cereal brand that you love and you remember that you saw the ad on TV or that that brand sponsored Raptors.

00;05;07;16 - 00;05;30;01
Speaker 2
And we watched the game and you're pulling on my arm in the grocery store and saying, Hey, I want to buy that. What I'm doing is helping brands know. Was that a smart decision? Should they continue to make those investments or should maybe we try to put the dollars somewhere else to reach the objectives? So actually, we have this conversation quite a bit with the kids and they love seeing advertising now.

00;05;30;06 - 00;05;34;18
Speaker 2
So my high school, I broadcast TV because of the ads.

00;05;35;14 - 00;05;37;07
Speaker 1
Like Mom, what do you think of this one?

00;05;37;21 - 00;05;38;19
Speaker 2
Yeah, exactly.

00;05;39;14 - 00;06;06;12
Speaker 1
Okay. So that is a really nice sort of simplified way of of explaining what this sort of art and science says. Can you tell me how you do it? Because as the as the marketer, I'm like, wow, that sounds amazing and would be so much easier. So sort of walk me through a typical engagement. You don't have to talk about who the client is, but what's a typical engagement look like with a client?

00;06;07;09 - 00;06;33;15
Speaker 2
Sure. So you have to start with the objections of the advertising media mix modeling or marketing mixed modeling. We can evaluate the impact on sales or on revenue. You can also run techniques to evaluate the impact on upper funnel metrics like awareness and consideration. So the first thing in an engagement is understanding what was your objective of the ad campaign or the the ad budget?

00;06;33;15 - 00;07;07;06
Speaker 2
Now for most consumer packaged goods brand sales, is that primary KPI that we want to evaluate and whether marketing was profitable. But if we're considering the B2B space or even an industry like automotive, the impact of advertising to lead to consideration is hugely important. So maybe the study is designed looking at that way. So we start with the objective and then we have to collect all of the data and there's a lot of data out there.

00;07;07;06 - 00;07;45;15
Speaker 2
So we work with partners across all of the digital platforms. We will work with partners who collect sales data. So whether that is a Nielsen or an IRA or even one of the digital e-commerce, this year we became a partner of Amazon to include Amazon sales data in our models. And so we will have all of that sales data, then we will get all of the media data, and that would include your traditional television, radio, print out of home, all of the digital buys, the programmatic buys, but the specific like direct buys on certain platforms.

00;07;45;15 - 00;08;06;16
Speaker 2
So like tech talks, you know, become coming up more and more now we see that your screenplays Snapchat Pinterest along with the ones that have been around for over a decade Google and mirror. And for some clients, there will be Twitter page investment as well. And we'll collect all of that data and then we have to harmonize it.

00;08;06;16 - 00;08;26;26
Speaker 2
So basically, like one of the reasons I wanted to call the company Mix Analytics is because we're mixing up this like big recipe of data and we have to harmonize it so that the data can talk to each other. So if we have unstructured data or structured data, you know, structured data is a big word because each data provider has a different structure.

00;08;27;01 - 00;08;56;23
Speaker 2
So we actually have to restructure it all so that it can talk to each other. And that's half the project. More than half the battle is getting the data, getting the right data, checking it and making sure it's accurate. Once you have that data, then the fun begins with the analysis and we will run multiple models exploring many different techniques to find, you know, a statistically valid product in the end that can help make those informed decisions for clients.

00;08;57;01 - 00;09;31;22
Speaker 2
So how much of your volume overall that will be typically delivered in a pie chart? So what percentage of that overall volume or that overall consideration is attributed back to the investments in each channel versus base sales or based consideration, which is essentially conversion that happened in the absence of any marketing? Okay. Brands that have great brand equity in the market, one of the values of brand equity is sales that consumers will buy that product without any marketing at all.

00;09;32;23 - 00;10;09;14
Speaker 2
You know, try to find ads for Honda in Canada right now and Slim because Honda's built such a fantastic brand equity that the cars on the road speak for themselves, that that brand hasn't been investing as heavily as some of the the other brands that are growing share in the market like Kiya, Volkswagen and others. So it's it's really that, you know, percentage of volume where we start and then we can drill into that.

00;10;09;14 - 00;10;40;08
Speaker 2
So the the mascot of Max is the little succulent and that's representing all of the layers that we peel back to evaluate marketing. And the layer of sales is one, but then the return on investment is this so your question? I invest in $10 million in marketing. Am I getting $10 million back? And so if we understand the amount of volume attributed to the marketing, we can then assign a financial value to that.

00;10;40;08 - 00;10;49;22
Speaker 2
Is that is that the ability of that volume or is of the revenue and divide that simply by the spend and you get your ROI?

00;10;49;22 - 00;11;26;06
Speaker 1
Okay. Yeah, that's very, very clear. I'm wondering I'm just wondering, I guess if there's any learnings that you guys have. So I know you work typically with consumer brands, but are there any sort of learnings about different channels at different sort of stages of the customer journey that you guys have been involved or common pitfalls or, you know, common sort of learnings that someone listening to this could be like, okay, you can take that away and sort of apply it to the way that I'm spending my marketing budget.

00;11;26;06 - 00;11;54;19
Speaker 2
Why don't we chat about that for each of the main buckets of spend? So let's start with traditional advertising and first, let's start with B2C and then we can chat a bit about B2B. So for for B2C, so the business is advertising to consumers and they want to drive sales. Then we're seeing that the that traditional advertising still has a very important role for most brands.

00;11;54;19 - 00;12;09;15
Speaker 2
Now, these are generic comments I'm going to make. Every single brand should have its own analysis because it's creative, it's household penetration and brand love is going to be different and that's going to be the ultimate goal.

00;12;09;15 - 00;12;16;06
Speaker 1
And you don't want someone listening to this podcast and then just being like, Oh, Hillary told me to do it. Yeah.

00;12;17;00 - 00;12;41;09
Speaker 2
Yes, all my money on TV. So we have we are still seeing the role of traditional advertising. And when I say traditional, I'm not just talking about TV, I'm talking about TV, radio, out-of-home and PR. And most of PR now is digital. But PR public relations has been around for four decades. So this is still traditional advertising tools.

00;12;41;23 - 00;13;00;25
Speaker 2
And the reason being that they still have mass reach. So if you really want to reach most most consumers and you have a consumer brand, that is something that majority of consumers are going to purchase or a majority of households have that in the house. So the household penetration is high. Then you still want a mass reach vehicle.

00;13;01;17 - 00;13;37;06
Speaker 2
It's going to get you that awareness the fastest and you know, and help drive sales. So we still see the role of that in most clients are asking what's the optimal percentage of traditional versus digital? And that we should look at it on a brand basis. Also depends on the amount of budget you have because if you have a lower budget, so this is a lower awareness brand or a growing brand, we we have a smaller budget to begin with then we may not have enough to do the right thing on traditional and we would advise actually a full digital budget.

00;13;37;16 - 00;14;09;09
Speaker 2
But in most cases we are still seeing enough, enough dollars available to have a traditional buy. So that's that's, you know, first and foremost, next is is digital. Now I've been measuring the impact of digital sense in sales since the early 2000s. So we're talking about like 20 years of analysis on digital. I've noticed CPMs cost per thousand impressions changing and the trends changing by vehicle.

00;14;09;16 - 00;14;34;11
Speaker 2
In each model run, you see these these changes, you also see where it's being purchased. So clients will shift where they're buying programmatically. They might flip their agency every, every so many years. And so you see these changes and the platforms are coming up with different ad tactics. I think Amazon has over ten different ad formats that you can buy.

00;14;34;25 - 00;15;03;23
Speaker 2
So with all of that change, the important thing with digital is testing each arm of your plan and really understanding the potential it has for your brand to your consumer audience by by running these tests. And what I mean by test is they have a period where you are investing in Pinterest and you have the right campaign and have the right creative for the channel.

00;15;03;23 - 00;15;26;15
Speaker 2
It should be specific for the channel and you run it out at the recommended size that Pinterest is going to recommend. And your or our advertising agency of record is going to recommend. And then you come back and you evaluate how that performs. Now, I am a huge fan of integrated marketing. That's when we have more than one channel running the campaign.

00;15;26;27 - 00;15;55;26
Speaker 2
But the problem, the pitfall in marketing, mixed modeling, is if we have all of that marketing happening at the exact same time, the models simply cannot tell us how much volume is attributed to each channel. If the data is highly correlated. So for us to really understand which what digital is working better than others, it's running these series of pulse tests across different channels to then be able to come back and understand how effective that was.

00;15;55;26 - 00;16;02;22
Speaker 1
For your brand and do you would recommend doing them in isolation? So like not doing them at the same time when you're doing the experimental phase.

00;16;03;02 - 00;16;28;24
Speaker 2
Yes. Or in combination or select markets using true a campaign running like an integrated campaign running across the country. And then you can choose certain markets where you are going to experiment and get a read on that and kind of rotate it. So if you're in that strategy, you have to also set up an agile framework to measure it so that you get to the answer.

00;16;28;24 - 00;16;54;02
Speaker 2
But yeah, so for traditional, just to recap traditional, there's still a main a main role because of its reach potential. And with a digital, we are in a stage where we still need to be testing because the assortment of media channels continues to change. We are now going to be seeing ad units available. I don't actually know if they're available anymore, but in Netflix and in Disney.

00;16;54;05 - 00;17;05;17
Speaker 2
So how are you now going to try to experiment and get into that space and get a read on it where it's not mixed up with everything else in your brand plan?

00;17;05;17 - 00;17;15;14
Speaker 1
Okay, that's really, really interesting. And then when it comes so for B2C, that makes sense. Does it change for B2B?

00;17;15;14 - 00;17;37;16
Speaker 2
Well, for B2B, I think a lot of marketers still have amazing digital tools available where they can be looking at the in the attribution space. So I know we were going to chat about a little bit about this big snazzy word multi-touch attribution model. I think such a mouthful. It's a.

00;17;37;16 - 00;17;38;13
Speaker 1
Mouthful.

00;17;38;17 - 00;18;04;17
Speaker 2
And really that's talking about the individual's journey. So the the individual buyer's journey and exposure and eventually what was the last click that led to conversion in B2C, that the data is really expensive for us to have that, you know, you're going to have to pay a lot of money to have a sample of consumers digital exposure and purchase habits.

00;18;04;29 - 00;18;29;06
Speaker 2
But in B2B, there's still a lot of great data in your web analytics that you can rely on for those questions and for testing environments. So I would say that there is still a role that that that can play in the the higher level aggregate analysis of of marketing mix. For me to be the trouble is often the data.

00;18;29;19 - 00;18;53;24
Speaker 2
I'll have many calls with marketers who want to evaluate the returns, but they haven't been tracking any of their data for a set amount of time. And these are a regression is a time series analysis. So you really want to be going back a couple of years to evaluate that so that you have more data points and more data points means more degrees of freedom in the model.

00;18;53;24 - 00;19;24;05
Speaker 2
And how many of these marketing executions are what we call independent variables we can actually model? So in a in a B2B space, we usually have a limited data set to begin with at an aggregate level. And then if there's no history of the marketing executions, it's tough. So I think for for B2B to continue to rely on the analytics that you can have in-house and through the different sites that you're running your campaigns on.

00;19;24;13 - 00;19;36;29
Speaker 2
But then start thinking how can I track from if I could go retro, it's fine. But from today onwards, how can we start tracking the right data so that we can evaluate how each channel is performing at an aggregate level?

00;19;38;17 - 00;20;01;21
Speaker 1
Yeah, and I think that sort of like I think we chatted like a year ago and that was sort of the advice that you gave me when it came to setting up my tracking, because when I joined today, it was like a blank slate and you know, you needed to set everything up from scratch and as you were talking about that for B2B, I was thinking from a B2C perspective because I know you typically work more with consumer brands.

00;20;02;29 - 00;20;26;29
Speaker 1
If someone were to be in a position where they're starting to think, I'm sure it's not typically startups that you're working with, it's probably more well established companies because they're starting to think about like, okay, we have bigger budgets, we really need to get a sense now of, you know, we have bigger teams running campaigns. We really need to get a sense now of where our ROI is coming from, what what would you recommend for those brands?

00;20;27;27 - 00;20;42;13
Speaker 1
I'm just thinking like, what do they absolutely need to make sure as is set up or set in place so that you could some you or someone like you could sort of come in or your team could come in and, and make sure that you're accurately calculating that. R y.

00;20;43;11 - 00;20;56;27
Speaker 2
Yeah. For, for B to B And I recall that conversation we had last year, and I even think about it in terms of my own marketing budget for analytics like in a, in a way part of my role. Also CMO of mix analytics.

00;20;56;28 - 00;20;58;11
Speaker 1
Yeah.

00;20;58;11 - 00;21;28;08
Speaker 2
You know, how, how am I evaluating the, the investments that I'm making and it's setting up those KPIs at the beginning. So, you know, each of us are investing in in-person conferences this year and you know, often the KPI is the few conversations like I would like to go to to Marie in San Antonio, Texas in two weeks, and I would like to have two solid conversations for that investment.

00;21;28;08 - 00;22;04;01
Speaker 2
Yeah. So I'll be working hard, busy b trying to have those two quality conversations that could lead to a couple of projects. And for you it you'll have a different KPI, but having alignment with the leadership on what that KPI is and then tracking that and sometime it feels exhausting because your ultimate goal is achieving that success. But you also have to have the rigor of tracking the activity and the owned channels that we operate as B2B marketers.

00;22;04;22 - 00;22;35;08
Speaker 2
We can't often get that data later from one of the platforms. We have to track that ourselves. So what do I mean exactly? Every month I try to track my LinkedIn followers, I try to track what were the impressions or engagements I received on certain posts. And so that over time when mix analytics hits, you know, to its or to your birthday in January, all have two years of data.

00;22;35;08 - 00;22;44;07
Speaker 2
Then of that marketing activity I've done. But if I called up my friend Jeff over at LinkedIn said, Hey, can you help me get this data? I'd be like, No way.

00;22;45;02 - 00;22;45;13
Speaker 1
To.

00;22;45;23 - 00;23;16;28
Speaker 2
Know where you have to get that data yourself. You have to track it. And in so it is exhausting. It's manual same thing with matter you know, for for any of us who have a an Instagram business page for stories and for posts like it's great. It's an additional channel, but only the paid data for at a sizable level is available from their analytics teams to pull retro.

00;23;17;09 - 00;23;29;00
Speaker 2
For those of us doing B2B, we have to collect that data in a timely fashion. Otherwise you forget and it can. You and I remember what we posted in March of this year.

00;23;29;00 - 00;23;30;23
Speaker 1
No, no, absolutely not.

00;23;32;08 - 00;23;35;07
Speaker 2
Well, it was International Women's Day, so I do remember that. But.

00;23;35;07 - 00;23;56;26
Speaker 1
Okay, fair, fair. That's a highlight. Yeah, I know that. That's a very, very good point. I think with B to C, I'm also just wondering from a tracking perspective, is there more to track like our in terms of the different channels that are being used?

00;23;57;21 - 00;24;24;06
Speaker 2
My team would probably help me if I was like, Yeah, we're still missing data because then when clients case for model, we're running right now we have 96 independent variables, meaning 96 different marketing tactics and campaigns we're or sub campaigns like a creative within a campaign we're run over three years to support a brand that's a lot of day to day database.

00;24;24;18 - 00;24;26;00
Speaker 1
That's a lot of data.

00;24;26;04 - 00;24;50;19
Speaker 2
I feel like for B2C, we have a solid learning and maybe more of a pitfall that I'm concerned about in that space is when some media mix models are run in, they are in absence of enough data, they're being run on a small sample of data and not the full story, and that can lead to inaccurate conclusions or even running it with the right datasets.

00;24;50;19 - 00;25;21;04
Speaker 2
But on a limited market scope like National Canada, well write for us without the gos. We know that marketing is different and that consumers are different. State by state or province by province. So we should be at this point in 2022, we should be running models with geo granularity. The data digital data providers can give us all of the impressions and spend for for B2C at that level and the sales data is available at that level.

00;25;21;04 - 00;25;41;09
Speaker 2
So why wouldn't we be running the studies with granular data for for B to B? It's a it's a different story is that we're still in the infant stage of having data and collecting the data because it often just becomes this other element in a brand manager's job, or they just don't get to that item in their inbox.

00;25;41;09 - 00;26;03;12
Speaker 1
That and with the with that was such a good bit of advice. Is there anything else that you would say to someone who's maybe already running like something like marketing, mix modeling, like maybe a consumer brand who's already running this? Any other watch outs that you could tell the listeners about?

00;26;03;22 - 00;26;27;09
Speaker 2
I don't know if it's if it's necessary. Lea a watch older, more like an area that I'm really excited about right now is that as the model softwares have improved over time and we've become smarter as a group of data scientists to run these models that we're now running studies at an item level. And so not just the total brand.

00;26;27;17 - 00;26;54;25
Speaker 2
Like if we imagine like a matrix and you have all of your provinces and columns and then you have all of your, your, your speakers and your items rolling up to total. We can run these models at such a detail level to then get the insights as to what, what is contributing to the overall profitability and to replace like an assortment analysis which is going to tell a brand which items they should have in the portfolio or which ones should have shelf facings and not.

00;26;54;25 - 00;27;18;11
Speaker 2
But it does tell us, like of that marketing campaign, what were the items or the subbrands that really provided the returns? And, you know, some of our clients that have seasonal items, if they were running a certain platform buy on digital or a campaign at a certain seasonal period. And a lot of it was coming from the seasonal volume.

00;27;18;11 - 00;27;42;15
Speaker 2
Well, if they if they decided in the future to run a forecast and say, I'm going to do that same campaign, but this time I want to run it in a different season. Well, you won't get the same return because that profitable seasonal item is no longer available on brands. So it's actually like advising them in a smarter way, like a more precise plan for the next year, considering all of those item rotations.

00;27;42;22 - 00;28;09;02
Speaker 2
And like you guys at Digg kind an upside are constantly looking at innovations and clients want to launch innovations well why would you want to know how effective the marketing is on the innovations versus the rest of the business? And the way we can do that is by having these models run at the item level. So I'm super excited, but I don't know if you can hear the passion.

00;28;09;02 - 00;28;30;12
Speaker 1
But yes, for anyone who's not actually watching the video of this and who's just listening to the recording, Hilary, as soon as she started talking about this, I went right into the screen. She got right up close. She was very excited, but that is incredibly cool and that's exactly why I was thinking and that's the exact use case I was thinking about.

00;28;30;13 - 00;28;55;02
Speaker 1
You know, what we do here is we help clients develop innovations, often with product innovation. So the idea that you could run any of this analysis at an item level would be so helpful for that. Our whips, you know, we can track sort of a lot of things post-launch, but that business of sort of tracking the ROI, that's something that obviously you guys excel on.

00;28;55;10 - 00;29;03;08
Speaker 1
And I think being able to do that as an item level is, is really, really cool. I didn't realize that you could do that. So very excited.

00;29;03;14 - 00;29;27;06
Speaker 2
It's so cool. And you know, a lot of brands have that general question like what's the impact of master brand or multi-brand advertising versus if I have a hero item in my in my it answers that question too, which is this age old direct versus halo like what was the direct impact of this creative supporting, this direct innovation?

00;29;27;06 - 00;29;51;09
Speaker 2
What was the halo to the rest of the items? Or if I have messed around, it would be in theory, direct and everything. But is it is it actually driving volume on every item? So this idea, this detailed model is, is super exciting for me and it and then when I see the results of it, it really makes me never want to go back to modeling at a total brand level ever again.

00;29;51;27 - 00;29;53;06
Speaker 1
So that's something that I'm on.

00;29;53;18 - 00;29;57;20
Speaker 2
The road now. I don't ever want to model total brand for.

00;29;57;27 - 00;30;00;25
Speaker 1
Explain more. Explain more to me. Why?

00;30;01;25 - 00;30;25;20
Speaker 2
Because. Because the results are there. It gets all lost. It's Nielsen taught me the term aggregate data bias. So when you look at data in aggregate, it's missing. All of the granularity can be hiding these trends that are within, you know, the item or the sub brand level. So yeah, I, I promise every year never went back to total.

00;30;26;00 - 00;30;33;13
Speaker 1
Will love it but that'll be the title of the episode. No More Total Brand Love of Love. Yes.

00;30;33;22 - 00;30;34;08
Speaker 2
Never going to.

00;30;34;08 - 00;30;52;20
Speaker 1
Think I'm conscious of time. I want to dove in a little bit to the fact that, you know, we are moving into a cookie this world. And I want to get your $0.02 on what that means for how your company or how you sort of think about attribution.

00;30;54;02 - 00;31;29;13
Speaker 2
So like I have said it, it's it's tough and it has been tough in Canada, very expensive to run attribution modeling for B2C and you know, because of the cost of data but also privacy laws. So there is a resurgence of interest in MMR, you know, even before going cookie less just frankly because empty multi-touch attribution is a way to hide in this market but but definitely cookie less is contributing to that and I think we just have to do it smarter and better than it was done before.

00;31;29;13 - 00;31;55;06
Speaker 2
For any brands who have maybe navigated away from I'm a mom and are coming back, it's where where I challenge you to really think through how to get the granularity that we can now versus going so granular down to the individual consumer journey level, but actually to that, you know, in a population base like some of the amazing Amazon data is available at an FSC level.

00;31;55;06 - 00;32;01;01
Speaker 2
So that's three digit postcode level. So you know, pretty good geo targeting beyond like total province.

00;32;01;17 - 00;32;01;29
Speaker 1
Right?

00;32;02;09 - 00;32;27;23
Speaker 2
So at that level and understanding the trends, you know, just so you're not going so deep and that's really going to be going to be the space and why I felt like it was a good time to launch a business and I'm a man because yeah, I know that the question just keeps coming up more and more, but clients are looking to do more with less always and do more in less time.

00;32;27;23 - 00;32;56;20
Speaker 2
So yes, where can we be agile and be affordable to answer maybe more questions? Relying on the machine, learning that we've built over time in these models. And and so you can answer more questions like the item level or in an agile framework. It basically means setting up your model and being able to run it, say, every month, every quarter, you know, maybe every six months, definitely annually.

00;32;56;20 - 00;33;24;08
Speaker 2
You know, years ago I used to tell clients in Canada was it was okay to run a model every two years. I think since we're still living in this post-pandemic time and we're seeing continued pressure on supply chains, the inflation interest rates that the consumer's purchasing power and consumer decisions are still going to be shifting across industries over the next couple of years.

00;33;24;15 - 00;33;37;00
Speaker 2
So I'd be very wary to use the results from two years ago to project into the future. I would say at a minimum, be looking to run these this analysis annually on your brand's program.

00;33;37;00 - 00;33;50;07
Speaker 1
This has been fascinating. Thank you, Hilary. We've got a couple rapid fire questions to close out. So first question, if you gained double your budget tomorrow, what would you spend it on?

00;33;51;27 - 00;33;58;00
Speaker 2
Okay. So for me, as with my CMO, Amex, I would spend it on more in-person conferences.

00;33;59;00 - 00;34;07;13
Speaker 1
Very cool. And what if you lost half your budget? What would you spend it on? Sorry? What would you what would have to go, not what would you spend it on?

00;34;08;10 - 00;34;16;25
Speaker 2
It's so tough. It's like holiday gifts or interesting conferences. I think I would cut the in-person conferences and I would keep the holiday gifts.

00;34;17;09 - 00;34;19;22
Speaker 1
Okay. Holiday gifts for your clients?

00;34;20;13 - 00;34;25;20
Speaker 2
Yeah. I'm way too attached to things like our little our little Lego kits and stuff.

00;34;25;26 - 00;34;27;15
Speaker 1
So cute.

00;34;28;04 - 00;34;29;24
Speaker 2
Too. Coming out soon.

00;34;30;15 - 00;34;44;17
Speaker 1
Oh, very exciting. And then finally, just to close out, is there any you've given a lot of advice today, really useful tactical advice, but is there anything else you wanted to leave the listeners with?

00;34;44;17 - 00;35;07;14
Speaker 2
Taste test the contest. Don't be afraid to test and test smartly, because like I mentioned with the correlations, which means basically marketing all happening at the same time with the same campaign across many channels. You know, that's a fantastic sounds like a fantastic integrated marketing campaign. Great. But we're not going to really be able to test the impact.

00;35;07;14 - 00;35;30;27
Speaker 2
So you have to set this marketing up smartly, considering how the math will run. And if you're not sure of it, then just call somebody like me as an expert and we can sit down and talk about how to build that brand campaign so that it can be measured because you sort of missed the opportunity of understanding the ROI in new channels and with new investments.

00;35;30;27 - 00;35;59;13
Speaker 2
And, you know, the we're not sure still the impact of campaigns on TikTok for for brands over time. So things like the ad stock impact, which is if you execute a campaign today, what's the decay rate, how many weeks later will it still drive a potential sale or build awareness or consideration? We need more time. Evaluating increases in like what's the saturation level, where how high is too high to be investing in a channel?

00;35;59;22 - 00;36;12;00
Speaker 2
And there's been great learning on Google and matter of Snapchat because we've been modeling that data for like decades, but a lot of brands are just getting into tech talk. So you want to.

00;36;12;00 - 00;36;12;11
Speaker 1
Be.

00;36;12;26 - 00;36;26;04
Speaker 2
During that detail, detailed result with your provider as well so that you don't either overspend in a specific week or underspend may be below the threshold where you're really not getting any impact at all.

00;36;27;02 - 00;36;32;04
Speaker 1
What a way to end. Thank you, Hillary. You're a legend. I will talk to you soon.

00;36;32;27 - 00;36;39;28
Speaker 2
This was great. Thank you so much for the opportunity and having the discussion and listening to me nerd out on all of my thoughts.

00;36;40;23 - 00;36;44;03
Speaker 1
It was awesome. I'll talk to you very soon.

00;36;44;03 - 00;36;47;21
Speaker 2
You take care.

00;36;47;21 - 00;36;54;19
Speaker 1
Thanks for tuning in this week. Find us on LinkedIn at Digg Insights and don't forget to hit subscribe for a weekly dose of fresh content.