When it comes to understanding your customers, one question looms large: Why can’t we predict customer behavior accurately? In a world overflowing with data, many businesses struggle to connect the dots between what the data tells them and the actual behaviors of their customers. In this post, we’ll explore the complexities of predicting customer behavior, including insights from AI-driven analysis, and why human intuition still plays a critical role in this process.
Key Takeaways
- Predicting customer behavior is challenging due to the disconnect between data and real-life actions.
- AI can enhance data analysis but should not replace human intuition.
- Transparency and ethical data practices are essential for building customer trust.
- Combining AI insights with human empathy is crucial for a holistic understanding of customers.
Transcript
Kchristaihne Castillo (00:01.55)
Hey everyone, and welcome back to #SoldOutChat
I’m K, the Marketing Coordinator here at Above Promotions. And today, we are diving into one of the most frustrating challenges for any executive, “Why you can’t predict customer behavior”. We’re exploring the disconnect between the data we collect and the actual actions your customers take. So to help us navigate this, we have our CEO, Ebony Vaz. Ebony’s background in engineering and data-driven research allows her to bridge the gap between complex systems and human storytelling.
Hi Ebony!
Ebony Vaz (00:36.54)
Hello, good to be back. Took a little break, but now we’re back at it.
Kchristaihne Castillo (00:40.91)
Yes, I’m really excited about this so let’s jump into our questions to help our listeners move from guessing to strategy. we often use the word predict but is it truly possible to predict a human or are we just performing highly informed anticipation? How does AI-driven data analysis fundamentally differ from traditional market research in this regard?
Ebony Vaz (01:08.562)
Yeah, that’s a good question. Think predict is such a loaded word. I mean, we know that if someone walked into a dark room, we could predict that they’re going to look, with most certainly look for a light switch to turn the lights on. And so that’s what we really do with the data that we have. We try to take a look at some of the typical things that people
do, how they perform in their elements, whether it’s in their personal lives or in their business lives. And then we attempt to see what is going to trigger them to do something. So that could be that whatever, and I talk about wants versus needs, what’s a vitamin versus what’s prescription medicine. And it’s kind of the same thing is there’s going to be times where we can predict
really relatively well what someone’s going to need at a certain part of their life and their journey. And so that’s what we really try to do when we talk about predicting humans and what they’re going to perform and what they’re going to want. And that is when we go to different websites and we think about some of larger websites that people order from all the time that they can kind of predict what someone might be wanting to purchase.
at a given period of time, maybe you typically purchase vitamins at the beginning of the month. And so they’re going to automatically send you reminders about vitamins that you can purchase, or they can predict that most people are going to be thinking about setting goals in January. So there are certain things that because as humans, we all often do, we can kind of figure out what’s going to happen. And so when you ask that question about how does it differ with
using AI versus traditional market research, the difference is speed. Yeah, the difference is the number of datasets we can gain access to on a very quick basis. And so to me, that’s really what is the big difference is AI has given us the speed, it’s giving us access to larger datasets and giving us the ability to parse that data much quicker than we could manually before.
Kchristaihne Castillo (03:35.37)
I agree. So, uhm.. with machine learning, Machine learning identifies patterns in vast datasets, but does it offer a glimpse into what a customer will do or just a likelihood of what they might do? Where should businesses draw the line between following an AI suggestion and using their own intuition?
Ebony Vaz (03:57.672)
So, yeah, I want to dive onto that for a minute because what can it predict what a customer will do? Again, this kind of goes back to like in certain situations, we know that if someone hears a fire alarm in a building and they are not expecting it, they’re going to move quickly through. And so there are certain things that we can kind of
ascertain that someone is going to most likely do. And so definitely machine learning has that ability to go through the data sets and make those decisions. We’ve used different levels of that when it comes to computer vision, where we’ve been able to do market research using wifi cameras. so based off of the sentiment of people’s face, whether they were frowning or whether they’re happy, know, those different types of things, it allows us to indicate, well,
this person could be excited enough to purchase something. They could be interested enough to want to learn more. so there is this way that machine learning can really predict what someone might do. And some certainties is going to predict what they will do. Now, go to that question and dive a bit more when we say where should businesses draw the line? So this is where
AI is in general is still so far behind because and let me let me back up for a minute. It is behind in ways that people are making decisions thinking whatever Gemini tells me, whatever Chat GPT tells me, whatever Claude tells me is the answer to this question. I can run with it. I can take this information and I can apply it to my business.
and I don’t have any questions for it. That is not the case. So even if you are building your own knowledge library where your AI agent can go and pull the information out, there’s limitations to what it will know. It will give you a high percentage answer off of a generic question. As you begin to ask it for more specific things, it can be more specific.
Ebony Vaz (06:21.914)
And as you start to drill down, and this is why, you know, you can recall when Chat GPT really kind of took off and was more public domain, people were saying, we’re looking for prompt engineers. And I think a lot of people did not realize the value in it because in that value, you need to know how to prompt. And so it wasn’t just a matter of having prompt engineering and knowing how to ask questions, but you need to have
people who are specialized in certain areas. Because if you’re trying to use this for, let’s create a marketing plan and I’m going to go to Chat GPT and tell it to give me a 90 day marketing plan. It could give you a marketing plan, but it may not know that during the third week of the fourth month of the year, no one purchases anything. It may not take into consideration religious changes. It may not take into consideration economic changes.
If the administration decides to change a law, it will not take that into account. It’s going to give you what it thinks to be highly probable for a generic answer. And until you dig down, and I mean dig, because there will be times where you can prompt and you come to a conclusion and you think, you know what, I like this idea it came back with, or I like this research it came back with. And then you might need to challenge it. You might need to say, now.
this idea, how would it compare or stand up against FTC law or how would it stand up against GDPR? You know, those are the types of things that you’re going to have to still know and you can’t just get rid of your workforce. And that’s why I kind of chuckle when corporations say, we are doing these layoffs because we don’t need these individuals because of AI.
I don’t think that that’s really what’s happening. I think they’re using AI as an excuse to lay those layoffs because everyone knows that you have to still have that expert in that field running certain questions. So that way it knows what to suggest. And so business owners still have to know their stuff. Department heads still need to know their stuff. We can’t allow brain rot and assume that the AI chat agent that you’re using is generating the right answer.
Kchristaihne Castillo (08:51.758)
So I’d like to dive more into when you mentioned the FTC law, GDPR, the privacy stuff. using AI and machine learning to anticipate needs raises questions about privacy and manipulation. So how can a responsible business ensure they are being transparent and ethical, avoiding practices that could damage the customer trust we’ve worked so hard to build?
Ebony Vaz (09:20.392)
So, mean, when we talk about trends and how AI is affecting those trends and can we really determine a trend or not, I can tell you without certainty, anytime some major corporation gets dinged by FCC, FTC, letter of the alphabet that the government regulates, whether it’s here in the United States or abroad,
Other corporations all of a sudden are like, are we good here? Let’s just double check and make sure we’re good here. So that’s another way that let’s say you are someone who offers services and compliance. You know that anytime there’s a breach or something, all of a sudden people need those types of services. And so we can think about building that customer trust and using these different types of tools, whether it’s AI and machine learning, computer vision.
any of those versions of artificial intelligence that you want to take a look at. We know it takes a lot to build customer’s trust. We also recognize that there’s levels to putting in your data sets into certain types of models and knowing that you need to be able to restrict either the information you put into that model or you need to have your own and build your own model. And so that’s going to be like.
The biggest crust of it is I know for us internally, whenever we are looking at new tools, we’re testing out new tools, whether it’s for our clients or internally for us, we’re looking at what the privacy says. If we’re collecting photos for a client, all right, what does this tool say they do with the photos for the client? Are they going to use it to train some model? Are the images going to be sold and used on some stock image site? Those are things that you really need to understand.
And also be transparent with your customers. Let them know like, Hey, we are utilizing these types of tools. These are the things that, and how it could affect you. And these are the parameters that we’ve put around these tools that we’re using. And so those are the things that are going to be very important in being transparent. I mean, we talk about it all the time. The critical part of the sales process is getting people to trust you. And if you can share with them how.
Ebony Vaz (11:44.304)
you are able to be trustworthy. If you can share your trust center, like we have a trust center on our website, tells people exactly how we intend to handle data for us, for our clients, et cetera. People want to know that. Give them a way to break through the barrier of deciding if they want to hire your company to provide whatever product or service that you’re providing.
Kchristaihne Castillo (12:08.77)
definitely agree with that one because especially with us, whenever we check these new tools, we ensure that we’re always checking that privacy part. So it’s very important that when we deal with customers and we have these third-party tools that we’re using, we definitely have to be transparent with them. So customer behavior is influenced by emotions and unpredictable life events that spread
So how can businesses combine insights obtained from AI or ML with human empathy to create a truly holistic view of the customer?
Ebony Vaz (12:50.834)
So I think again, this kind of goes back to you need to have some level of expertise within your organization. Even if you are hiring someone new, you either want them to bring that expertise or you’re going to provide that expertise inside of them because of the fact that AI is only going to know so much. There are years of lines of information that you have read and done.
that’s coded in your own brain that will not be coded into AI until you determine to put it in there. And so it’s going to be extremely important that we understand that there’s this human element that’s needed. So again, I say the example, if a corporation got breached today, I guarantee you, CISOs, general councils for other organizations are like, okay, let’s double check and make sure this is good.
yeah, when was the last time we talked to our HVAC vendor? When was the last time that we made sure that our secretary knew what to say and not to say? So those are the things that will definitely influence behavior and also ways that we need to be cognizant of that because we know if something happens externally and it affects our industry, we know that that could affect the selling cycle. We know that that, you know, how that could generate other
triggers for potential customers or existing customers. And because of that, we have to be able to add those insights to what we get. Yeah, we can go to GPT today and exit. What is the historical trade information for cotton being imported into the US? We can get that type of information. But what it may not know is perhaps
that there is now a trend that people only want to use grass instead of cotton as a material. That is something that’s going to be missed. And so this is why we have to take a look at what we know versus what the data knows. And let’s say there’s this movement again from cotton to grass as the medium of choice to make clothes. We have to be cognizant of
Ebony Vaz (15:17.682)
What are the reasons why? Maybe all of sudden the places that are bringing cotton into the US aren’t matching with the values of people who are purchasing their products or perhaps something terribly wrong with the cotton coming in from other countries or even within their own country. So you have to really have an understanding of what are customers thinking about right now. And that’s why we’re going to always need this human element. That is also why you’re seeing more people
Being okay, I remember a time where if you had a humanities degree, if you had art degree, you really were getting passed over. But now your organization’s looking for that because they need people who have anthropology degrees, people who can understand and try to understand the humans more and what they do and their thought patterns more. And those things are, it’s going to take a minute for AI to be able to adopt and move.
in a manner that makes sense for human interaction and human decisions that need to be made for businesses.
Kchristaihne Castillo (16:26.008)
Actually, funny thing, I just recently watched a Korean movie and it’s called… Yeah. So I watched this Korean movie and then it talks about this certain person, she’s like the lead in that movie and then she’s being saved by that certain government or company because her…
Ebony Vaz (16:33.114)
you did.
Kchristaihne Castillo (16:54.13)
ability or her skill is to create the emotion engine. This is to teach the new humans. it’s like even in the movies they’re already showing that human element is really important in teaching this AI. So
Ebony Vaz (17:10.747)
It is.
Kchristaihne Castillo (17:13.262)
Behavior isn’t static, it really evolves. So how can businesses use AI not just to predict current actions but to anticipate major shifts in market trends before they happen?
Ebony Vaz (17:25.894)
So I think it’s, can they predict current actions? So again, we have to look at historically how do people interact when they have some type of trigger to them. And so this is where it’s like, is it really predicting or is it just anticipating? And so we can expect that
If there is a trade change and trade policy, it’s going to affect businesses in different ways. If the trade tariffs go up, that affects them. If they go down, that affects them. So we understand that there’s going to be shifts in the market, especially if the market is being turned around or changed course by certain laws that may happen.
So we have a good understanding about that. And so we can start to anticipate that if there’s too many shakeups at once, we can anticipate that we’re going to have to anticipate changing again. Sometimes it is this part of understanding that if we’re in a place of a lot of volatile change that we know that we have to be ready to pivot. And so what that does, it means that we have to start to
get more, maybe more team members, or we have to make sure we stock more of certain items, or make sure we expand certain software, because we know we’re in this place that we have to be able to scale or compress very quickly. And so we can use AI to say, if we have to compress really quickly, what do we need to do? If we have to scale really quickly, what do we need to do? And I hate to say it, and I am pretty sure that people
are using AI to determine when to downsize in workforce and when to upsize in workforce. And that’s a part of where you hope in all of this, there’s human element and it’s being taken into consideration for these decisions, but it’s happening. And so we know that we can use AI to make these decisions in different manners. It’s like,
Ebony Vaz (19:44.168)
We know usually in the springtime, we can expect people to eat a little bit more healthier. So grocery stores are gonna be ensuring that the front of the store has all those healthy brands that people know and love and some new ones in there to try those out. And they know as the seasons move on to like the summertime, we know people are thinking about the barbecues, the family vacations. And so we see this thing that happens to shift. And so…
grocery stores, they’re using AI to determine like where they’re at with the inventory levels. And so there’s different ways in which it can be utilized to count certain things, personnel, people, items, products. We know that because of that, and we have this history of data, we can anticipate like what it can do. You know, we constantly look in and marketing.
And when you have social media campaigns, you’re always looking to see what the analytics look like. And so we can use AI to kind of predict that. But it doesn’t mean it’s going to be right. I mean, we’ve seen it. And one of our tools that we use for social media, it will say, AI is predicting these are the good times to post. And then we post at that time, and those were not the good times to post. So it’s not like AI is this magic tool. We still have to.
measure and use the data. And so that’s, think for us, no matter what we might be current focusing on as even as our business, we’re always going to fall back on this storytelling science and technology part because it all fits in together. It’s all integrated. We’re always having to constantly.
take a look at the story, seeing how we have to adjust it. We have to take a look at how consumers are interacting with it from that data science point, and then what technologies and tools we’re using to implement it and when we’re implementing it at. And while we can anticipate that AI can help us in that in the future, it’s not fully there yet.
Kchristaihne Castillo (21:51.318)
I definitely agree with that one. As you mentioned, human element is really important and if we did not use that, we’ll definitely doing worse than we have been doing. So good thing we still have that human element and we don’t let go of that. So finally,
Ebony Vaz (22:02.254)
Yes. Yes.
Ebony Vaz (22:09.605)
Yes.
Kchristaihne Castillo (22:10.946)
How should businesses measure the success of their predictive tools? Like, what specific metrics should they track to ensure that these AI efforts are translating into tangible business outcomes and revenue?
Ebony Vaz (22:24.272)
Yeah, so I think one of the things is we have you have to stick with some basics. The basics is you don’t go and add a bunch of tools at once. You just don’t go and do that. It’s like building a house. You don’t just tell the guys go ahead, start pouring the foundation while you tell the guys to go ahead and start putting the walls up at the same time.
It doesn’t, it doesn’t work. doesn’t make sense. So you have to really be methodical about this. And so if you want success, you got to start with a tool at a time. See a tool implemented, you know, contact us, let us, you know, we can give you some of our recommendations on some of the tools. can do a tool audit for you and see what is working now and what’s not, but it’s really best to try making small changes one at a time and not trying to do everything.
on that once because that’s when things break. And that’s also when you are not able to determine if what you’ve implemented is actually working. So I really think that businesses in order to be successful, you need to start one at a time. And then you need to determine like what are the metrics we’re going to be using to determine if it is working. So if we just take social media from one, there’s vanity metrics like, I have 1 million followers, but
one million followers and we go to a post and only 200 people interact with it.
You know, those are, it should be counting how many followers you have or how many people are actually interacting with it. And I think that’s, those are the type of things we have to like determine, like what are going to be the actual KPIs? What are the things we’re going to measure and what is most important? So perhaps at some point in time, the company is coming new onto the scene. What is the first thing they’re measuring? Well, they want to be the first thing to measure is converting leads, but
Ebony Vaz (24:18.374)
really probably the first thing they need to measure is, is their brand getting recognized with that traction. So you have to set your metrics based off of what your current goal is and be realistic about that goal. Having those smart goals in place because making sure that you’re setting metrics that are relevant and you have their time base and time bound are going to be extremely important because you don’t want to sit there and keep a tool for like
20 months and realize no one in the company is using the Or realize this tool is costing you money and it’s doing the exact same thing another tool that you’ve already paying for is doing. So this is going to be very important for you to ascertain like which goal is most important for the company and the organization and help that to help you drive what metrics is set. And it doesn’t matter if it’s AI tool. It doesn’t matter what tool it is, but you have to have some type of metrics.
put into place and you have to be smart about what you decide to put into place as well because there might be a tool that does a lot of the work that you want to do, but it might violate HIPAA regulations or it might violate privacy laws. So there’s things that you really have to take in consideration. We’re always, always for us in terms of any tool governance and compliance is number one. That is number one for us. We obviously
making sure it gives you and performs what you want to do. That is good. And always good in theory, because a lot of software will tell you in theory, does X, Y, Z, but before you even go down that rabbit hole, find out where that data is being stored. Where is it being processed? How is it being processed? How do you have access to remove that data? Cause we talked before in a couple of podcasts about how you need to make sure you remove those.
tools that you’re not using anymore and what happens to your data, et cetera. So there’s a lot of things that go into selecting the right tools and making sure that you’re auditing them on a regular basis. And it’s one of those things that people really overlook and it’s those things that usually get you in trouble. But we’re here, BufferOcean is here to help you out. So whenever you are in need, please let us know. Let us come in and do that audit and let you know what could be.
Ebony Vaz (26:45.18)
effective for you because if you are that you own the company or not, if there’s going to be some type of transaction where that company gets sold, et cetera, they’re going to be looking at what tools did you pick? What tools are in the company? How are you using those tools? Are you just overloaded with the number of software to conduct your tools? And there’s a whole list of things which we’ll get into in another podcast coming up soon. But yeah, when it comes to measuring and predictive tools,
We’ve got you covered.
Kchristaihne Castillo (27:19.682)
Thank you so much, Ebony, and thank you for helping us understand how to make customer behavior more predictable. And to recap, we discussed shifting from reactive marketing to informed anticipation. While AI provides the “What?” through vast data patterns and probabilities, the “Why?” remains a human endeavor. So check out our other episodes on Spotify or YouTube, and be sure to connect with Ebony Vaz on LinkedIn. Thanks for tuning in to #SoldOutChat and stay strategic,
and keep winning.
Conclusion
In conclusion, predicting customer behavior is a multifaceted challenge that requires more than just data analysis. By understanding the limitations of traditional market research, valuing human intuition, and fostering trust through transparency, businesses can navigate this complex landscape. Moving forward, the best strategies will be those that embrace the strengths of both AI and human empathy.
While you can’t truly “predict” a human, you can resolve the technical and strategic friction that makes their behavior seem erratic. At Above Promotions, we specialize in identifying the hidden variables that cause these disconnects.
Whether through a comprehensive MarTech Audit in our Innovation Lab to smooth out the digital journey, or deep behavioral analysis in our Strategic Insights Center to understand the “why” behind the data, we help you move from reactive guessing to highly informed anticipation. We help you find the holes in your strategy and turn the mystery of “unpredictable behavior” into a clear, scalable, and intentional path toward growth.
Ready to find the holes in your strategy?
- Watch the full episode on YouTube here
- Book a Consultation with our Innovation Lab to audit your brand alignment and stay ahead of the curve.
Check out some of our favorite items the team uses to make our recordings work:
– USB to HDMI adapters https://amzn.to/4aFkoQv
– Tripod https://amzn.to/4kImvHO
– Carpet Cord Covers https://amzn.to/4kImvHO
– Reusable Cable Ties https://amzn.to/3MCzsGu
– Wireless Lapel Mic https://amzn.to/4rVD4lV
