So What? Marketing Analytics and Insights Live
airs every Thursday at 1 pm EST.
You can watch on YouTube Live. Be sure to subscribe and follow so you never miss an episode!
In this week’s episode of So What? we focus on Holiday Predictive Forecast. We walk through how the holidays look for different industries, where to source predictive forecasting data, and what actions to take with your predictive forecast. Catch the replay here:
In this episode you’ll learn:
- how the holidays look for different industries
- where to source predictive forecasting data
- what actions to take with your predictive forecast
Upcoming Episodes:
- TBD
Have a question or topic you’d like to see us cover? Reach out here: https://www.trustinsights.ai/resources/so-what-the-marketing-analytics-and-insights-show/
AI-Generated Transcript:
Katie Robbert 0:27
Well, hi, everyone. Welcome to so what the marketing analytics and insights live show. I’m Katie joined by Chris and John, we are now in that weird in between time of the last couple of months of the year where Thanksgiving has happened. Christmas has not yet happened or any other real major religious holidays, and everyone is just sort of like, alright, let’s just get through this. And so with that, John, John and I were talking the other day about, well, does every industry sort of follow the same kind of holiday predictive, like holiday seasonality? And the answer? I guess we’re gonna find out today. So what we wanted to chat about, was that holiday predictive forecast. And so what holidays? How the holidays look different for different industries? Where to source predictive forecasting data from? And Chris, you’re certainly going to touch on that, and then what actions to take. So Chris, where do you want to start today, as we’re talking about predictive forecasts,
Christopher Penn 1:33
I guess, one of the more important things to start with would be to understand what the purpose of the forecast is. So Katie, when you put together user stories, what would you say would be an appropriate user story for, say, a, a cheese shop? Let’s go with that to start, you know, what would have predicted forecasts for chi shop there? What would their user story be?
Katie Robbert 1:53
So as the manager of a chi shop, I want to understand what she’s people are looking for on any given week, so that I know what to be promoting on social media have recipes for and then a second one might be as The Cheese Shop Manager, I want to know, what cheeses are people not looking for. So I know what to put on sale and clearance. Manager Special.
Christopher Penn 2:27
Special shoes is blue, and it’s not supposed to be
Katie Robbert 2:32
special does mean that things are pretty much out of code. So Buyer beware.
John Wall 2:39
Avoid the manager special.
Christopher Penn 2:41
It’s like saying you always wash your cans before you open them. Okay, so with those stories, that’s a great place to start. Because one of the things you want to do is you want to use the various different tools that are available to marketers to see like, what kinds of trend data could we get if we were to go out and get set data? So let’s look, let’s go to our two of our friends, right? Let’s go to friend number one is going to be Google Trends, right, which is arguably one of the most useful pieces of software that is still available for free. And what this does, for folks who are unfamiliar, is it you simply type in search terms, you can see the historical data going as far back as Google knows, for tribes. So let’s go ahead and start with mozzarella, right? Because that’s a cheese. And we’re in the United States, which can change location, we’re the past, well, let’s look at the last five years. That way we can see, you know, are there any kinds of ebbs and flows in this. And so if you’re a cheese shop manager, certainly just being able to look backwards and see what kind of things things kind of happened here. And you can see, there are definitely some, you know, a kind of a heartbeat cycle like a sine wave within the data. So there is there is something of a trend here. So this would be one of the tools that you’d wanted to get started with, just to try and understand like, what can we even predict trend? Right can because Predictive analytics doesn’t work with things that are unpredictable. It seems kind of obvious, but people forget that part. It’s like, everyone’s trying to predict the 2024 presidential election, you can’t predict something that hasn’t happened. So
Katie Robbert 4:35
right and looking at historical events. It’s not a one to one match. So you’re not looking at the right kinds of predictive metrics because it needs to be something that is consistent and reliable. And you know, for a lot of reasons presidential elections are not those and they’re different every time different people different, you know, circumstances variables. So I think that that’s a really good pro tip to keep repeating.
Christopher Penn 5:05
Exactly. So that’s part one. Part two is using the SEO tool of your choice, we use h refs, but pretty much any tools got keyword stuff in it can can do this. Put in all of your search terms, and just get a sense of, you know, this data is the last 30 days, most tools default to the last 30. And you can see the number of searches. By the way, John, I don’t know if you hit mute or not, but that did not mute.
Katie Robbert 5:32
Just gonna say the same.
Christopher Penn 5:38
But you can see the dip the volume, all that if Jesus so in the last 30 days, we see here Wahaca cheese, cotija cheese, mascarpone, cottage cheese, cream cheese, and so on and so forth. Lots and lots of all these different uses. And we see the the number of searches. So people searched for search for cream cheese 94,000 times in the last 30 days, they search for GriefShare and 91,000 times the last 30 days. They search for Parmesan cheese. So again, it Katie, if you were The Cheese Shop Manager, this would at least give you a starting point, the challenges. This is all data that looks backwards, right the last 30 days, which as everyone knows, tastes change, you know, circumstances change, we’re about to go into the winter holiday season. And there are probably cheeses that you would serve at Christmas that you wouldn’t necessarily serve other times the year. Like it was always that one that comes in the like the sausage and cheese basket that looks like it looks like a spreadable cheddar cheese. But then it’s got like nuts and stuff on the outside. It’s like that a cheese log roll shrink wrap thing. Yeah, yeah, cheese balls, Hillshire farms and stuff sells those things. And I mean, again, that’s not stuff that you’re ever going to serve outside of the holiday season. But when it when someone gives you the Hillshire farms box to Harry and David box, you’re like, oh, there’s there’s the cheese ball with stuff on the outside.
John Wall 7:02
Thank you very little.
Christopher Penn 7:05
So all that says is we want to be able to look forwards, not necessarily backwards, because circumstances can change, particularly when you’re dealing with consumers. With B2B Things are a little more cyclical because of things like stock markets and quarterly reports and how budgets work and makes things a little bit less volatile. Okay, so those were the first two data sources, I would start with that. The challenge here now is we have to kind of glue these things together, and then build forecasts from them. The way we do that is with a type of software, predictive analytics software, using an algorithm called profit with x g boost. This is time series forecasting Software released for free as open source software from Facebook, of all companies, and then modified by the developer community to add error correction that Facebook for whatever reason didn’t include in their software.
Katie Robbert 8:11
Well, you know, it’s funny, I remember, gosh, it must have been six, seven years ago now, when you rolled this out to our then marketing team. And I remember my first thought was, this is a Facebook thing. And I think everyone on the team was also sort of confused of like, what does Facebook have to do with this. And so knowing that the developer community has taken this open source code and modified it to be more accurate is super helpful.
Christopher Penn 8:42
Yep. Facebook uses this a lot for their advertising. And the reason why it is valuable is the methods they chose incorporate a lot of seasonality, but they incorporate a lot of weighting towards more recent trends. Obviously, if you were selling advertising space and an app, you want to you want to give more preference and weight to forecasts based on recent data than legacy data. Because, again, things change, people change. If you think of a service like Instagram, there may be a new hashtag or a new challenge or a new trend that appears overnight. And you don’t want to be trying to forecast and do advertising bid pricing on trends from even seven days ago, because you want to take advantage of this new trend like eating Tide Pods.
So the catch with software like this is as you can see, it’s available a couple of different programming languages are in Python. If you don’t code in either of those languages, there’s not really a way to use the software. So there are products that do this out of the box and IBM Watson is an example. But those obviously do come with a bit of a price tag.
Katie Robbert 10:07
Well, I would think, too, they’re also harder to customize for, you know what you need or just more difficult to understand how the algorithm is actually working versus building it yourself.
Christopher Penn 10:19
Yes. And one of the things that not enough tools do. And we actually had to write this ourselves in our own software is to say, do a statistical check to see if there’s a trend before you write a forecast for it, right, because we don’t want to try and forecast something that doesn’t exist. So let’s do forecast for me. We’ll write the app. You can see over the last five years that you’re basically there’s some mistakes, some typos, but it wasn’t until really. Yeah, March of this year, that this app really took off. And they they got the Saturday Night Live placement and all that stuff. There’s not a trend here, right, we see a shape that sort of looks a little like a bell curve, but from a statistical perspective, there is no cyclicality. There is no seasonality. So this cannot be forecasted. Because it’s sort of a once in a lifetime thing. Really good example. Oh, god, oh, clubhouse,
Katie Robbert 11:31
I was gonna say, wasn’t there that thing? I think I was gonna call it chat house, which is definitely not what it is. Right?
Christopher Penn 11:39
It definitely is not what it is. And what you see is pretty clearly Yeah, there was there was sort of that big spike, it sort of came and went. Not a trend. Now, by definition, which means that you cannot forecast from it. However, things like mozzarella cheese or Parmesan cheese. These are things that absolutely do have seasonality and cyclicality let’s remove you can see even with parmesan cheese if I remove the mozzarella here there is that that heartbeat rhythm so you would your next step would be to feed this into predictive analytic software and get a forecast of some kind get an output and since we’re talking about let’s start with the cheese once it’s talked about different industries let’s go clear this
and switch to our cheese database cheese she’s gonna put it in our dates, our terms, our numbers, and let’s take a look just at 2023. And let’s turn this into a line chart a term our color and bring this up. And so we get Yes, well for each of these cheeses, now we can look at at these cheeses either in bulk or or individually. You saw the in bulk with the Trump look like spaghetti. And I’m saying okay, here’s you know, here’s when in the next year American cheeses like the peak, right? Right around July, 2 peak around November another peak a December and then a peak in early January. So you have these these ebbs and flows in American cheese. You could take have already Yeah, let’s go to have already have already a bit different Thanksgiving and Christmas but you don’t have that summertime bump. Right? It’s it’s not as popular cheese. If we look at mozzarella cheese. Mozzarella, you get a decent summertime bump. Why, of course people are cooking in their pizza ovens that that they have outdoor.
Katie Robbert 14:20
Well, it’s also mozzarella goes really well. And it’s what the mozzarella tomato and basil salad. Caprice, a summer salad. Thank you. So you know, it’s interesting, because you know, you’re doing the whole United States, which is a really great place to start, especially if you have like an online business. And so I would say that, you know, if you are like a brick and mortar store, drilling down into just your state is also going to be really helpful, which you can do through Google Trends, because the trends that you see for the entire United States might be different than the trends You see for just your state. So, you know, you, when you’re talking about things like food and like the seasons and seasonality, the way in which the weather happens in your state is going to impact that. So in Massachusetts in July, and August, the humidity is off the charts. And people don’t want to turn on their ovens and you know, sort of sit inside with like this hot hot food. So they’re likely to eat more of the crazy salads and other things that are easy to prepare outside. And so you might see that trend different from, you know, Alaska, for example, which is in the United States, but the weather is different at that same period of time.
Christopher Penn 15:44
Exactly. So that’s if you were a sort of a cheese person, if you were in the architecture, industry, architecture, engineering, and construction, again, very similar. Let’s go ahead and take our term and come to a series of pages here. So people search for architects, at very different times of the year, right. So people search for architects, you know, early in January, mid April, we see August, and then outs outside of it. So those during those times, periods of time. That’s what people are looking for. That when people are looking for engineering firms. They don’t
Katie Robbert 16:22
well, probably like structural engineering.
Christopher Penn 16:25
Exactly. Yes. So so I’m looking for structural engineer near me. Here we have again, January March, that we have May, July, August, September. So again, sort of which goes against what you would think for a normal B2B, normal B2B Like office work. We know that people are out of the office during those times, you know, they’re on summer vacation, but people searching for structural engineers, that’s actually a summertime thing. Why that is? I don’t know, because I don’t I’m not a structural engineer. But I would imagine it probably has to do with budgets and project timelines.
Katie Robbert 17:00
Well, you know, if you think about, you know, what you need structural engineers for? I would personally, if I were, you know, in this space, I would be looking for, you know, what are the trends of structural engineer near me versus contractors near me, because those tend to go hand in hand. So for example, I needed to look at, I needed to have a structural engineer to look at the back of something on my house, before I could have some work done because of the type of landscape that I have. And so typically, when you look for one, you’re looking for the other. So if I were a contractor, I would want to know when people are looking for structural engineers, and vice versa. And that would also sort of tell you, especially again, depending on what part of the country you’re in, you know, contractors tend to be busier in New England during the spring, summer and fall months winters just, it’s tough to do that work, versus someplace where it’s perpetually summer like California.
Christopher Penn 18:02
Exactly. And one of the interesting things in this forecast is that the overall macro trend is on the increase, right? If you look for just even simply eyeballing it, you can see that in terms of popularity, this search is increasing in popularity over a long period of time. So if you’re a structural engineer, or working in a structural engineering firm, this might be a very good thing for you and your business. So that’s something that you want to take into account with trends and forecasts, as well.
Katie Robbert 18:30
So I think one of the things that we wanted to demonstrate is that, depending on the industry, the trend is likely going to be different. So not everyone is going to have the same holiday rush. That sort of we’ve come in no for the retail industry. So as we’re seeing with structural engineers and architects, there is no winter holiday rush per se it’s more of a spring summer thing. Cheese depending on the cheese, the type of cheese, it’s definitely going to vary between summer and winter. You have American cheese, which is really popular at summer barbecues, hotdogs, hamburgers, you know, all of that good stuff that goes on the grill versus in December. American cheese isn’t as popular because maybe you’re not grilling outside as much. And you’re wanting more of those like festive like aged cheeses. I’m making that up. I’m lactose intolerant.
Christopher Penn 19:32
Really good example is vacation rental. So this is a very popular term right. Vacation Rental that season. If you were if people start searching right around spring break right now right on the April vacation week, just after Easter. And so in 2023, mid late April is when people are really going to start looking for that and that peaks the searches for it peak, a sort of end of June. Soon after that’s on the backside. And that is pretty obviously like people trying to find where they’re going to go on vacation. Right. But you see, there’s a couple of early peaks in January, and in February as well. So, you know, those may correlate to, to school vacation weeks, again, looking for combinations. And again, towards the very end December, we see people who are looking for vacation rentals for the holiday season, but again, a very, very different look. Very different industrial look, then the other two.
Katie Robbert 20:33
John, what do you think?
John Wall 20:36
Yeah, well, it’s just, that’s the most important thing to keep in mind is that for your industry, you’ve got to have the data to be able to see these kinds of curves. Because if you just think it’s going to be the classic seasonal thing of it’s going to be quiet in December, and nothing’s gonna happen in the summer, you can totally get destroyed. And we see this too, because it’s, you know, we are always talking about how we think July, August, September will be dead. But for us, that’s for our customers, that’s prime planning time, or they’re actually free to go into other projects. So what’s normally a sleepy time and B2B, you know, in other fronts ends up being really busy for us. So, yeah, you, you know, if you don’t have this data, you’re just kind of flying blind. And there’s a lot of bias with this, too. You know, there’s people that think, when things should be big or not big and you know, only the data can really get you on the right track.
Christopher Penn 21:27
Exactly. All the things you can forecast, we’ve not done this here, because it requires a bit more lifting as you can for essentially, if you’ve got any data that goes by time, you know, cashier, cashier, register receipts, shopping carts online, you Google Analytics data, your email marketing data, your terrestrial radio, ad impression data is long as dates, and there’s, there’s numbers. And it’s regular and frequent, you can forecast it. And that’s, that’s the important thing to remember about these technologies is they are agnostic, you don’t just have to use Google Trends, you can use any time series data. And the thing about forecasting is like everything else, the more specific it is, the better. So if it’s your Google Analytics data, or your Adobe analytics data, or your Hubspot data, you’re going to get a better forecast that’s unique to your business than taking an industry wide forecast.
Katie Robbert 22:31
Exactly. You know, Google Trends is a really great place to start. So we that’s where we started, when we first launched Trust Insights, we didn’t have that historical data. So we were relying upon systems like Google Trends and SEO tools to tell us what people were searching for. But now that we have our own, you know, five years, basically now worth of that data. Our goal is to be using our own data, to see how people are searching for us and what kinds of things they’re finding us for, and then use that data. And so it’s a constant evolution. Because what we see in our own data might differ from what Google Trends was telling us.
Christopher Penn 23:14
Exactly right. And the thing that we have to keep in mind, too, is that you may be in a situation where, yeah, your data is a hot mess that so don’t rule out search data, even if things are a little bit on the messy side, because you might be able to use public search data. If, if your data is just a disaster, let’s take a look at management consulting. So management consulting, this looks like traditional B2B. Right, so you have your first quarter, lots of interests you have sort of your second course ag you have planning, there’s your planting season people, you know, August, September, and then you have you sort of just fall off the cliff here after Thanksgiving. So that’s, that is classic B2B. Very easy to forecast. So you can see, you know, they each industry has its own rhythms.
Katie Robbert 24:17
Well, and I think this so what of that is, you know, if you are just making assumptions about the seasonality based on what traditional, you know, consumerism looks like, you have your holiday season, you have your major, you know, things like Valentine’s Day and that kind of thing, that may not be the seasonality that works for your business. And that’s why doing this kind of forecast specific to at least your industry is important to understand, you know, those peaks and valleys and so now I think we’re at the point of like, alright, we have all this data, what do we do with it? What can you do with this information?
Christopher Penn 24:57
Right, and that’s, and that’s the question is, it depends Hands on what your user story is that your user story should be telling you. This, for example, is Google Analytics 4. You know, it’s, it’s on a pretty clear direction as to people’s interest in it. But there’s some big spikes that are going to be happening in sort of April. And then again, in July, which we know, the July one is when Universal Analytics just turns off, on stuff. But we also know that that April spike is probably going to be a bunch of people going,
Katie Robbert 25:30
I only have one quarter left to do it.
Christopher Penn 25:32
Exactly. So based on the user story, if we go back to the cheese shop, example now, let’s actually duplicate this and remove our term and go into our tabular data. Let’s say you’re the cheese shop owner. And it is, let’s do the week of Valentine’s Day. Cheese and love are in the air.
Katie Robbert 26:01
I mean, who am I to judge?
Christopher Penn 26:03
Exactly, provolone cheese, all by pecorino fall by Swiss, those the cheeses that are going to be most popular that particular week, in February. So as you’re planning out your Valentine’s Day theme, and your store and your point of sale displays and all that stuff, those are probably be the problem peccary not supposed to be the ones you’ve have upfront. And you have to decide as a as a business person, based on how you know people use those cheeses, whether you should be doing a discount and a sale or raising prices, because demand is going to be there.
Katie Robbert 26:39
Now, how far ahead do you feel is, you know, say for someone to be putting that information out there. So obviously, it’s going to differ by the type of channel that you’re using. But so let’s say we know around Valentine’s Day that provolone, you know, is going to be the thing. Should I start promoting Provolone for Valentine’s Day in December? Is that too far ahead? Should I start doing it at least two weeks out? Or is it doing it just that week? You know, good enough.
Christopher Penn 27:14
Oh, that’s where we’re looking at the data and see like, what you see in each of these data points is sort of inflection points where things bounce, generally speaking, you want to promote just after the inflection point. So here for provolone, there’s a soft spot in in May, right, that’s where that search volume kind of ends, then the following week, the end of May, you go on this upward trend as search volume goes up. So we have to remember the data source to this is people searching, which means they have intent. Someone is looking for stuff about provolone. Increasingly, until you get to sort of this this mid June thing. So if you were that store owner, as you know, search intent is increasing, that’s when you probably start putting out maybe some digital mailers, maybe some emails, some social posts, asking people, Hey, what’s your favorite provolone cheese recipe, you know, things like that. And then as you get sort of to that, that the midpoint of that curve, or even, you know, earlier, early to midpoint, start running keyword based ads and pay per click advertising and on social media, where keywords are part of the ad targeting. When you get to this top, this peak here, that may not be when people are going to use the cheese now with cheese, that’s probably the case because it’s a perishable whereas something like vacation is going to have a more of a time lag. That’s when you could say okay, yeah, maybe we can start to back off are our most expensive paid channels, for example, because we’re know that that demand is is going to start to gather but that’s where I would say, you don’t want to look just at that week, you want to get the context for that term with those basket of terms and say, Okay, here’s, here’s what’s happening. And when I should get started versus when we know everybody’s going to be in in the knife fight.
Katie Robbert 29:03
Make sense? I’m still trying to figure out like what’s romantic about provolone cheese, but that’s where I’m stuck.
Christopher Penn 29:12
provolone. This is wildly guessing because cheese
Katie Robbert 29:18
expert I love wild guesses. So let’s prove loads
Christopher Penn 29:21
of soft cheese that you can melt you can melt with other cheese and make some great fondues.
Katie Robbert 29:25
Ah, again, lactose intolerant, I wouldn’t know that.
Christopher Penn 29:31
You can also use it on pizzas. It’s not as good. It’s kind of a weird cheese. Smoked Provolone is very good. That you can eat by itself.
John Wall 29:41
What would you think about adspend? Like my gut would be you definitely want to be dumping spending when you hit the inflection point when search volume is going up. Would you actually turn it off or dial it down? Once you hit the next inflection point and it starts to fall off. I mean, in theory,
Christopher Penn 29:56
it depends. It depends on how much demand there is and how much time Competition you’ve got one of the advantages of predictive analytics is that, like, if you know, you’re going into a major season, let’s go to our heels go to will be good. Inbound Marketing, right? If you know that you’ve got a major competitor, say like Hubspot, that’s gonna be, you know, just flooding, places with money, you might want to go just at the bottom of the inflection point, just that early start, because you might be ahead of them. Because most people who go by gut and assumption go don’t have the exact, you know, days or weeks. So you could maybe get a market opportunity by going a week ahead of a competitor and get some really sure. setting the expectation to that because search volume isn’t as high, you might not get as much performance, but you might get folks who are who are starting to think about it, but haven’t started searching for it yet. But you know, it’s going to be on their mind.
John Wall 31:02
Yeah, that’s funny, you told you see the inbound show there in November, right? That’s
Christopher Penn 31:07
Yep. I would say, with anything paid, you have to take into account competition. Because and, and also, the other thing you have to take into account is inventory. Inventory is an issue to this past year. If you were doing Holiday Gift Guide advertising, you are paying a fortune, in September, October for your for your ads, not because other people were doing that. But because the inventory is being consumed by politicians running ads, they were just flooding every ad network with as much money as they possibly could. Everywhere they could, for all the different elections in the midterms. And that just take me off, that’s a soaks up inventory. And so everybody’s advertising gets more expensive when you’re in an environment like that. Now, thankfully, for those who are forecasting 2023, you don’t have to worry about that. But in two more years in 2024, you know, that until you know, whatever, November 7 of that year, there’s going to be increasingly less inventory during a presidential election in the in the United States and your ad your ad budgets have to be accounted for. And that’s, I think a really important type of predictive analytics is being able to look at the macro picture. And understand, okay, from a macro picture, here’s what’s happening, that’s going to affect everything, not just certain key words, but everything. When we look, for example, at the Federal Reserve Bank’s 10 year to three month Treasury rate, which we’ve talked about on on the show in the past, that indicator, when it goes below the zero line, means that investors have very, very low confidence in the market. And a recession usually occurs within a quarter. You can see, as of this morning, when I ran this, this is fairly significantly red. So investors are very, very bearish right now on the market. And that also means that companies will be spending less so again, if you’re setting ad budgets and stuff. It hasn’t happened already, you will probably get the call saying yeah, we’re cutting our budgets for 2023. And you’re like, we just talked to a client yesterday. We talked to them three weeks ago, like oh, yeah, we want to do this what this talk to them yesterday, oh, our budgets got cut, like, Okay, here’s why.
Katie Robbert 33:25
So it sounds like you know, the big sort of takeaways, the bullet points are, you know, if you are looking at trying to understand seasonality, make sure that you are comparing yourself to the right industries, the right companies, if you don’t have enough of your own data, there are data sources that you can use, such as Google Trends, you can use SEO tools. Those are really great places to start in terms of understanding consumer behavior, your customers behavior of how they’re searching for you. And then as you are getting more of your own data, start to transition into using that data as a supplement to the organic search data, because it’s, it’s going to give you more granularity into your specific customer base.
Christopher Penn 34:14
Exactly right. And keep an eye on on the macro picture. Keep an eye on the big picture, the things you’re going to be influencing stuff in the long term because we are we are perpetually and ever increasingly living in a more unpredictable world. And we will be over the next 50 years, right. We saw over the last three years, you know, Hey, someone parked a ship the wrong way and the Suez Canal, international trade stops for 14 days and pandemics and billionaires buying social networks and all sorts of crazy things have been happening. That unpredictability means that your forecasts have to be taken with a grain of salt understanding like yeah, these are great Rate tools for for making plans. But be ready to just hit the fling the forecast out the window and say, We gotta go with what’s happening right now. Because this happened and nobody’s you know, this was not on anyone’s bingo card.
Katie Robbert 35:12
Well, and I think that that lends itself to running a forecast, you know, for a year a year out, is okay for planning. But you’re better off doing it in smaller periods of time and running it more frequently as things that are now in the past, like the beginning of the show is now in the past, it’s already happened. And so that’s not something we can predict anymore. Because it’s gone. It’s, you know, bye bye. So running your predictive forecast, once a month, once a quarter, depending on how reliant you are on it for planning is a really good idea. Running it once and saying great, this is the plan. It’s never going to change. Not a great idea.
Christopher Penn 35:55
Yeah, think about like the weather. How often do you check the weather forecast? Right? Yeah, well look at the 10 day forecast, like two weeks, but you check the weather every day, like you say, okay, in two weeks, I’m planning as big vacation because the 10 day forecast looks good. And then you never check the weather again. Until then you’re like never your day like oh, hey, look, now it says it’s gonna rain now says great tornadoes. Who knows. But all forecasting is like that there’s there is uncertainty. And in environments of uncertainty, it’s a really good idea to check the forecast more than once more than just once you would not. I would hope, just pull a weather forecast for a year in advance and say, Okay, I know exactly what to wear every day. I’m like, oh, that’s gonna go for you.
Katie Robbert 36:41
Is that how you use the weather forecast? John, you look at it once a year and just gonna say YOLO.
John Wall 36:46
Right. Yeah, exactly. There’s there’s no bad weather. It’s just bad clothes, right? How about if I want to get some of these reports? I mean, I don’t have time for this. I’m busy running a cheese shop. I don’t have time to write reports, where do I get this stuff?
Christopher Penn 37:03
If you have the resources to do so obviously, we do them clearly. If you don’t have the reports, because they can be reassuringly expensive. There, you can still look at historical data. And if you know your industry, well, you can make some decent guesses, right? If you’re a cheese shop, and you look at the last five years worth of cheese data, and you see trends in it, you can eyeball it will have less accuracy, it will be less accurate, but you can say okay, you notice it looks like every MAE Provolone starts to pick up. So I know at some point, I should probably be doing something a Provolone sooner rather than later. And so in those instances, if you are willing to accept less accuracy, which we just talked about on the podcast this week, that is that is unacceptable. You cannot afford the full data driven machine learning power forecast.
Katie Robbert 38:02
But if you can, you can contact John Wall directly. He will take your calls. John stuff. Yeah, John’s, here he is.
John Wall 38:15
A mirror camera getting you there.
Christopher Penn 38:17
It really is. And again, the macro stuff. Everyone should be paying attention to that right. I generally suggest to people that they read news sources like Reuters, or AP news, places that don’t have a political spin, just like here’s the news. Because knowing what’s going on like, hey, in the early days of the pandemic, India, certainly a couple of big population centers in India were heavily affected. Those population centers make the precursor ingredients for acetaminophen. So if you know your industry, and you know how things work, so as supply chains work, when that population center gets shut down, you know that three months down the road, there’s going to be a Tylenol shortage, because of just the way everything is interconnected. So if you have a lot of domain expertise, if you know that this year, dairy herds were severely impacted by the drought, you know that, you know, cheese prices are going to go up. And if it’s an aged cheese, like an aged Parmesan cheese, it’s aged for three years or whatever, you know, that come you know, September October of 2025. There’s going to be a parmesan cheese shortage and you’re going to need to you should be stocking up your stocks. If you have the ability to carry a long term inventory like stuff in the deep freezer, and be expected you’re gonna have to charge higher prices and you may have customers who are impacted. So a good part of the macro of predictive analytics is knowing what’s happening right now. That’s going to have second order effects down the road.
John Wall 40:03
Alright, I’m good for the parmesan cheese shortage of 2025. We’re ready.
Christopher Penn 40:08
Put it in your bunker in the vault.
Katie Robbert 40:13
All right. Well, I think until next time, that has been your cheese talk.
Christopher Penn 40:20
Thanks, everyone. We will see you next week.
Unknown Speaker 40:25
Thanks for watching today. Be sure to subscribe to our show wherever you’re watching it. For more resources and to learn more, check out the Trust Insights podcast at trust insights.ai/t AI podcast and a weekly email newsletter at trust insights.ai/newsletter Got questions about what you saw in today’s episode? Join our free analytics for markers slack group at trust insights.ai/analytics for marketers See you next time.
Transcribed by https://otter.ai
|
Need help with your marketing AI and analytics? |
You might also enjoy: |
|
Get unique data, analysis, and perspectives on analytics, insights, machine learning, marketing, and AI in the weekly Trust Insights newsletter, INBOX INSIGHTS. Subscribe now for free; new issues every Wednesday! |
Want to learn more about data, analytics, and insights? Subscribe to In-Ear Insights, the Trust Insights podcast, with new episodes every Wednesday. |
Newsletter
Podcast
