Podcast with Dr. Wanda Curlee, Program Director, School of Business and
Dr. Tracy Fisher, Faculty Member, School of Business
Artificial intelligence, or AI, is constantly learning about user behavior online. How can businesses collect and use this information to improve their marketing strategy and reach new customers? In this episode, Dr. Wanda Curlee talks to APU business professor Dr. Tracy Fisher about how AI is incorporated into popular platforms like Facebook and Google and how business owners can leverage this technology. Learn why it’s especially important for small- to medium-size businesses to take the time to optimize data collected from these platforms, understand the analytics gathered, and how to incorporate it as part of their multi-layered digital marketing strategy.
Listen to the Episode:
Subscribe to Innovations in the Workplace
Apple Podcasts | Spotify | Google Podcasts
Read the Transcript:
Dr. Wanda Curlee: Welcome to the podcast, I’m your host, Wanda Curlee. Today we are going to be chatting about building an artificial intelligence, AI, strategy in business, Facebook and Google Analytics. My guest is Dr. Tracy Fisher, who is a part-time professor in the business administration program at American Public University.
Welcome back to the podcast, for our listeners I will remind you about Dr. Fisher’s background. She is a veteran of the telecom industry. She was in the industry when Internet of things first appeared on the market, which includes smart refrigerator, smart thermostats, smart phones.
[Podcast: How to Start a Career in Artificial Intelligence]
Unfortunately, she was in the industry during the dot-com crash. Dr. Fisher worked for two companies that did not make it to the other side of the crash. However, the network which was deployed during the boom is still in service, the transport layer is still in use, the boxes on the ends have been upgraded multiple times. Tracy, welcome to the podcast and thank you for joining me.
Dr. Tracy Fisher: Thank you, Dr. Curlee. I appreciate the opportunity to be a contributor to the university’s efforts in providing a world-class education to our students.
Dr. Wanda Curlee: Wonderful. Feel free to call me Wanda, please. Tracy, please remind us what your definition of AI is and why it is important to the business.
Start a Business degree at American Public University. |
Dr. Tracy Fisher: My definition of AI is very broad. A machine that can adapt to its environment, make decisions or problem solve and act on the decisions, learn from the outcome of the decisions and continue the process multiple times.
Dr. Wanda Curlee: Oh, interesting. So your background is a fit to analytics and AI; provide us some background on Google and Facebook business analytics.
Dr. Tracy Fisher: Well, recently I’ve helped a couple of local nonprofits set up Google websites and Facebook sites. Included in this process is setting up the analytics side so owners can create or adjust their digital marketing strategy.
I worked on small-scale nonprofits. Google or Facebook is a great place for nonprofits to start. The actual website or page on Facebook is free, there are some add-ons that can be purchased, typically they’re very inexpensive, such as a customized email address.
Overall, with so many web hosting sites and the cost of server space going down, even the pay sites are way down in pricing for web hosting. A decade ago it was a major investment for a small company to create a website and pay for SEO services, which is search engine optimization services. Now there are so many competing in this space you can host a basic website for less than $25 a year, or in some cases such as Google and Facebook, it’s free.
Dr. Wanda Curlee: Wow, that’s incredible to know, but why do businesses have to have an AI strategy for Google or Facebook analytics? Doesn’t Facebook drive the strategy?
Dr. Tracy Fisher: The data for Facebook shows the social media company has around 1.6 billion users per day with 2.8 billion overall user accounts. And the data for Google has an estimated 3.5 billion searches per day.
So just in a sheer volume approach, businesses using Facebook or Google Analytics have access to potentially millions of customer hits. If a business isn’t tracking hits, as in collecting the data and utilizing AI analysis or strategy, they’re missing opportunities to identify customer demographics.
So by collecting customer demographics, businesses can target advertisements on Facebook feeds or Google advertisements as in the popups. Google owns YouTube, so Google controls the pop-up advertisements on YouTube videos. So when you see those ads come up on Facebook, those are really in the background being pushed out by Google.
So does Facebook and Google drive the strategy? I would say no. Google and Facebook bring the customer to their multiple platforms. It’s up to the owner of the site or the owner of the business to drive their digital strategy. Facebook and Google are in business of moving Facebook and Google forward and profitable. It’s up to the business owner, using these platforms to drive their data analytics strategy.
Dr. Wanda Curlee: So by looking at their strategy or what they’re getting in pop-ups or what they’re getting traffic, then you’re saying that Facebook and Google don’t drive that, the actual business does, correct?
Dr. Tracy Fisher: Correct. The actual business is going to put into their search engine optimization a profile that will drive business to their site. If they’re not getting business to their site, then they need to look at what they have in their profile that they’re asking Google and Facebook to analyze and drive business to them, but it’s up to the business owner to create that profile.
Dr. Wanda Curlee: Okay, well that’s interesting. Who should drive and own the strategy then?
Dr. Tracy Fisher: This really depends on the size of the company. If you’re at a large, multinational, national corporation, these companies will have a fully staffed data analytics department or data analysis positions that are embedded within each department.
So it depends on how the company is structured, as in structures that are functional, divisional, flat or matrix. So it depends on how the company is structured. To give you an example of a functional, I worked in operations; if I needed any type of data, I had to fill out a form, identify the purpose or the strategy of the report or the data that I needed, all the fields I wanted in my report, and send it to the IT department. The IT department was a functional department within our organization.
Unfortunately, when you have this type of setup, results were not first in first out. The IT department or director or manager would prioritize the department personnel hours, they only have so many people with so many hours to build these reports. And it was not out of the ordinary for other departments to out-prioritize the data I was requesting.
However, now, currently most companies have deployed user-friendly internal software. This software is either homegrown within the company or it’s off-the-shelf like PeopleSoft, and departments can build their own internal reports, so the functionality has been moved from a separate department into individual departments.
Now, if you’re looking at medium- or small-size companies, what might be considered regional company, a local company, an e-commerce, or a hybrid, so they have a brick and mortar store and an online presence. The strategy at this point is owned by the sole proprietor or perhaps a partnership.
As with most decisions within this size company, decisions are made at the ownership level. That does not prohibit the owner from hiring a data analytics person or team, but the strategy is owned by the owners.
Dr. Wanda Curlee: It seems like it’s more critical at the medium or small that the business owner or whoever they hire for doing data analytics, or somebody within the company, understand how to drive or how to set up the web presence to drive consumers to that website.
Dr. Tracy Fisher: Yes, that’s correct.
Dr. Wanda Curlee: Okay. That could be difficult for small to medium businesses because they may not understand all of that, but hopefully through this we’ll find out how AI can help them, which leads me to my next question: How does AI fit into business Facebook and Google Analytics?
Dr. Tracy Fisher: Facebook and Google use AI in their business strategy. But let’s identify some keywords before we move on. The user is the individual with a Google or Facebook account. The term customer is the business who has purchased ad space within Google or Facebook. So that’s the Google Facebook customer, as in you and I are typically just Google and Facebook users.
A hit is a user click, which can be if you’re on Facebook and you want to like someone’s post, you have a couple of different emojis that you can select, so that’s considered a like. And then you have user reviews on both Facebook and Google. You have posts on your Google site, you could set up easily customer comments, which would be considered a post.
And then on Facebook, we all know that we can put posts in there. Here’s an interesting fact, if the user focused on a certain item for more than an average user, this is considered a hit. So, for instance, if I keep scrolling past an ad on Facebook, versus if I stalled out on an ad for a second or two to read it, this is logged as a hit. Facebook AI is so in tune with human behavior, AI knows the difference between just scrolling over it, or a view, but no click.
But back to the strategy question, Google and Facebook use AI to identify what item or user just hit. AI determines all the demographics associated with the relationship between the user and the hit, and this is where AI comes in.
By learning this relationship, AI will create a profile and correlate the user profile with Facebook and Google customers. And again, the customers are the companies that are buying ads. This is all done on mega-global servers. It is amazing, the vast network of processing power Facebook and Google own.
I did some quick statistics, looked up on Facebook, in 2019 is the latest data that they have out, but Facebook used 5.1 terawatt hours of energy. A lightning strike has 8,000 watt hours. So Facebook used the equivalent of 5.1 trillion lightning strikes in 2019.
Dr. Wanda Curlee: Oh my goodness, that is, wow. That was a great analogy there. Wow. That is amazing. Understanding all of this, what sort of analytics are possible from Google and Facebook analytics? Can these be changed for a company or do companies set them up?
Dr. Tracy Fisher: Well, all types of data are possible from Google and Facebook analytics. From a user’s perspective, this depends on what demographics and information users placed into their profile when they created their account with Facebook or Google. And what relationship the user has created within the platform, the multiple relationships that you have within Google and within Facebook.
As an example, what Facebook groups do you belong to? Who are your friends on Facebook? Who are your family members on Facebook and what groups are they in? What Facebook pages am I following? What information is in my post? Did I put my physical address, phone number, date of birth, all these relationship statuses into my profile? If you kept some of this information private or can only be seen by your friends, Facebook can see all of it, whether it’s out to the public or still private.
So if the only information placed into a user profile is a name and email address, which I think you had to put a minimum of an email address to create a Facebook user account, then you’re really at that point little use to Facebook as far as them being able to use AI to drive ads to you, because all they know is your name and your email address.
It isn’t until you start using Facebook, until you start making friends and making hits inside the platform, liking something, stalling over when you scroll through an ad, and so forth. At this point, AI can make decisions about you.
For instance, if you’re always liking posts on restaurants in a specific geographic area, AI pretty much learns where you are and that you like to eat out, so now they have two data points on you. For example, if I own an Italian restaurant nearby, I definitely want my business profile in the SEO to reach out to you in some fashion with an ad. Because I know that you like to eat out, you went to another Italian restaurant, you take the time to post reviews and so forth. AI is constantly learning everything about you as you interact on these platforms.
So as for the business side, it is similar to writing an abstract for a paper and having a list of keywords in the abstract. When a business creates a profile, it is saying to AI: When you see someone searching for anything related to Italian in this geographic area, send my information to the user in the form of an ad.
Now, if I decide to also want to expand my business model and incorporate “I sell Italian wines” I would need to either create a new profile under a wine related company, or I would add this keyword to my SEO profile and allow AI to identify when and where to place the ad.
Dr. Wanda Curlee: Well that’s interesting, but I do have a quick question around that. So when you say put the keyword into your profile to drive those people to your site, is this something that the average user can do? Or do I have to have a lot of knowledge on how to write this into my profile?
Dr. Tracy Fisher: Yes. You would need to add these keywords or something related to that in your search engine optimization profile.
Dr. Wanda Curlee: But is that something that, if I were a business owner of a small company, it would be intuitive to me as to what I needed to do, or should I hire a consultant?
Dr. Tracy Fisher: You could definitely hire a business analytics consultant. There’s actually a certification out there that individuals can obtain through Google Analytics and get an analytics certification. It isn’t very difficult to do, but it may be a daunting task for someone who isn’t technologically savvy, so I would recommend getting some advice from an expert.
Definitely you can go out to YouTube and there are plenty of YouTube videos on how to create a profile and how to optimize your SEO, so you can either be self-taught or you can seek advice.
Dr. Wanda Curlee: Okay, that makes sense. What would happen if a business does not drive an AI strategy for Google or Facebook business analytics?
Dr. Tracy Fisher: Well, again, this depends on the company size. A large company that hosts its own website and is on the internet directly, as opposed to using Facebook or Google, will probably continue to thrive. For example, Kohl’s hosts their own site on the internet. Rarely will you see Kohl’s advertisements show up in users’ feeds or Facebook users’ feed or Google YouTube ads pop up.
However, the Kohl’s website, their actual website probably gets a million, I’m just estimating here, users per month, maybe more. This is a lot of data. It would take AI or machine learning to collect, analyze, and provide a move-forward strategy. And these strategies can be developed by product, by region, by customer type for Kohl’s. But again, once you’re signed in to Google and you go to the Kohl’s website, Google will ship you all your data to Kohl’s. It’s in the agreement.
However, this isn’t to say that Kohl’s isn’t also collecting data on their Facebook page, they do have a Facebook page. But most customers are going to visit the official Kohl’s website to shop or use the Kohl’s app, Kohl’s has an app in the Play Store.
And what this shows is Kohl’s has a multi-digital marketing strategy across multiple platforms and is leveraging all the AI to determine who is the customer and what do they want, and at what price are they willing to pay?
So without using Google and Facebook analytics or AI, it is the medium to small, regional or very localized one–shop business that will miss an opportunity to introduce the business to users and identify the analytics by product, region, or customer. It could potentially be a major factor in the business going under as more people move toward the technology of finding the products that they want, locally.
There really is no viable reason that I can think of for a company that has a Google business or a Facebook site not to be taking advantage of this tool. While there may not be a lot of machine learning per se, as in Google and Facebook internal AI platforms, which are massive, there is still the opportunity for analytic strategy using AI tools on the Google or Facebook platform.
One of the driving factors for these two options is the price. Small businesses are very price conscious, they have small budgets. Google and Facebook platforms are free to create. It’s free to create your own Facebook page or a very, very simple Google business site.
Now there are some add-ons or up-sale cafeteria type items that you can purchase, customized emails for everyone in your business, or you could have a customized email for returns, for technical questions, for sales questions and then these emails can be directed toward a specific employee, internally.
I believe on Google SEO for the hits though, if you’re looking at purchasing ad space on Google, you only pay for if someone clicks on your profile. So if you pull up Google Maps and your Italian restaurant shows up, and a customer clicks on it, and then goes to your website, then you pay. If you’re on Google maps, your profile is in there, it shows up in a search, but they don’t click on it, they scroll past it, I don’t believe you have to pay for that.
Dr. Wanda Curlee: You said something earlier about having to go through your IT department to ask for different analytics. So should different departments within a corporation ask for different business Facebook or business Google Analytics?
Dr. Tracy Fisher: Yes, of course. Each department is going to have its own requirement for data to continue to support the overall business mission or goals or strategy. If we look at the different analytics between sales and return, so these are two obviously different departments. Marketing department would clearly be interested in what a user is viewing, what they’re searching for on their website, what are they placing in their cart and checkout, and what could potentially be upsold.
So for example, AI could determine that 40% of customer’s put item A in their cart, they also purchased item B, so you will see a recommendation. And we see this all the time as you go to checkout, you’ll see the recommendation, “Well users that bought this also preferred this.”
So marketing is moving products in front of the customer and collecting the data that everyone that bought A also bought B, C and D, so they bought more of B, so let’s recommend B when they buy A. So marketing would want to see that data and determine how to move that product forward.
Now, if you look at the return department, interestingly enough, the return department may not actually handle the return. The item could be sent to a wholesaler who purchases and packs up bins full of returns for various retailers and sells the bins to the highest bidder. Or basically the return is just shipped straight to the dumpster. And this is a big problem on returns.
Interesting data on returns is upwards of 30% of online purchases are returned. If you compare that to brick and mortar stores it’s less than 10% of purchases are returned to the store. So, right now, the industry is trying to figure out what is driving this difference in returns. The returning is very expensive to online retailers, it’s very difficult to pack those up and resell them once they’d been opened or perhaps used.
So back to the returns, the department that is responsible for returns and product quality will want the analytics on what items are being returned and why. So if you’re returning something online, typically you’re asked, “Why are you returning this?” And you’ll have either an option to type something in, or you’ll have a drop down list of most common answers like didn’t fit, purchased by mistake and so forth.
So return department will want this information, along with the data analytics that makes recommendations, to determine if they need to go back to a supplier because the quality is bad, or if they need to resize something and so forth. So return department, which is different than the marketing department, would want different analytics.
Dr. Wanda Curlee: That’s quite interesting. So you talked about marketing, you talked about returns, but should every department within a company, or every owner of a company, receive analytics? I’m just curious, are there just certain pinpoint organizations that should receive the analytics or should it be the leaders of each department?
Dr. Tracy Fisher: That’s a good question and again, it depends on the size of the company. If you’re a small business, obviously, it would be the owner of the business. If you’re like a medium or regional, you may have quasi-department that manages the analytics and sends it out to all the other departments.
If the retrieval and analyzing is too complex for a small business owner, you can, as I’ve mentioned earlier, hire a Google or Facebook certified analytics expert. For a business owner, Google has a program as I mentioned, that you could perhaps send one of your employees to this training and it’s all online. I don’t know if Facebook has this option available, but a small business owner can go to sites like Upwork and hire someone who is educated in AI and analytics to do the work.
So there are options for owners who do not see themselves as tech savvy, or just feel like I just want the recommendations in front of me, I don’t want to have to do the number crunching and so forth.
In larger organizations, anyone who has a responsibility of making or implementing strategic decisions for the company, I would let these people have access to the data, I would let them build their own reports and you can do this in PeopleSoft. These could be your managers, this could be a designated person within a department, depending on the type of data. Obviously if you’re a large organization and you’re publicly traded and so forth, you’d have to have disclosure agreements and so forth in place for whoever has access to certain types of data.
Dr. Wanda Curlee: I remember actually getting those kind of things that you’re warned during this time you can’t reveal this situation because of stakeholder or shareholder data.
But, let’s talk about the future; do you see AI tweaking and updating the strategy and the retrieved data based on what is happening on Facebook and Google and what the leaders are asking for? Or will this have to be programmed into the AI each time?
Dr. Tracy Fisher: I would say all of the above. Both. AI can quickly identify in real-time trends and present the data to decision makers if they haven’t already programmed into: If you see this, do this. But if it’s something else that is trending that might be related to this, AI will present this information to the decision makers.
However, this has to be coded into a business’s data or digital strategy. Again, AI needs some upfront knowledge before it can start learning. So for example, let’s say a video is trending on Facebook based on views and likes and shares. If the subject of this video is correlated to a business’s digital strategy identified in the business profile or SEO options, Facebook or Google can send notification to the appropriate person via whatever notification system that the company has placed in there. And businesses can make real time active decisions: Do we want to weigh in on this? Do we want to start putting ads out on this?
Another example would be Google searches. If the volume of Google searches for a specific toy goes up, and we see this a lot right before Christmas, this could be an event we’re all-too familiar with as parents, finding the most popular toy for our kids, the toy that every kid has to have. Businesses might want to optimize profits, and we’ve seen this before, and adjust the price of a product based on Google searches.
So if a lot of people are searching for this product and you have three of them, do you want to raise the price a little? And it’s just basic economics of supply and demand, we see this every year at Christmas.
Dr. Wanda Curlee: Yeah, I remember when SpaceX took off, there was actually I think, a purple monkey, or there was a purple stuffed animal that was seen floating around, and I got told that the sales for that purple stuffed animal just skyrocketed, so I’m sure they were taking some business analytics on that, whoever the company was, I have no idea who it was. But will those business professionals that do not have a working knowledge of AI be at a disadvantage?
Dr. Tracy Fisher: Yes, most definitely. I would not say all companies or professionals, because I’m sure there are some exceptions out there, but for the most part companies that do not implement really a multi-layered digital marketing strategy. And multi-layered, again, is having an app, having a Facebook page, having their Google site, a Twitter account, all of these things as a multi-layer digital marketing strategy. Using AI and data analytics technology, they will be at a disadvantage compared to a company that is leveraging all these tools across all the technology to bring customers to their sites.
Whether the business is a brick-and-mortar site or a complete e-commerce or somewhere in between, they will be at a disadvantage. Depending on the size of the company, every company needs to have either a functional or divisional analytics department. For small businesses either the owner needs to become data savvy or contract with a remote worker to analyze the data or train someone within and send them out for that Google certification.
Dr. Wanda Curlee: Let’s say a business owner really doesn’t want to do anything with Google or Facebook, for whatever reason, analytics. Are there other options that they can go to?
Dr. Tracy Fisher: Oh, absolutely, there are other games in town, definitely. Squarespace, Wix, GoDaddy, all the web hosting sites will have some form of analytics. Customer may have to pay to play, but there are options out there.
Dr. Wanda Curlee: That’s interesting to know. So there’s some competitors to Google and Facebook, although I’m sure they’re the biggest two in town. So Tracy, thank you very much for joining me today, this has been absolutely fascinating, analytics and AI to me go hand in hand. Do you have any final words to share with our listeners?
Dr. Tracy Fisher: Yes, thank you. My advice to our university students would be to learn all aspects of AI and how the technology is involved in the formation of business strategies. If after graduation you want to open a business, leverage the data from your multi-layer digital strategy to move your business towards customers. Again, your business is trying to find the customers. So take time to optimize the data collected from your Facebook page or your Google page or whoever you’re using to host your website.
If going into a professional corporate setting after graduation, whether multinational or regional, just determine the method of data transfer within the company. And leverage the information to support the company’s strategy within your department. If you have career aspirations, as we all do, always be preparing yourself, not for the next step, but for the step after.
Dr. Wanda Curlee: Tracy, thank you again. And thank you to our listeners for joining us today. Stay well.
Comments are closed.