In 2018, a common thread across Community Signal conversations was knowing how to clearly communicate the success of community across an organization. It’s fitting that for our first show of 2019, we’re joined by Tammy Armstrong, someone who is passionate about using data to solve problems.
Tammy and Patrick met on KarateForums.com and from her years as a community member and moderator, she learned a valuable lesson (care of Oprah): “When we know better, we do better.” This lesson carries throughout the entire episode. When we mature and become more empathetic, we become better community members. When we know more about our community goals and shortcomings, we can use the data at our disposal to do better. When we’re using the internet ourselves, we should be cognizant of the trade offs that we’re making with our data and whether or not it’s worth it.
Tammy also shares several tips and tools [22:30] for those that are just starting to unravel the data behind their online communities, but her greatest (and cheapest?) bit of advice involves understanding the goals and vision for your community. From there, data is just another tool to help you reach your desired outcome [40:36].
Tammy and Patrick also discuss:
- The positive impact of teenagers within online communities
- Best practices for data visualization
- How to start unpacking the wealth of data behind online communities
Our Podcast is Made Possible By…
If you enjoy our show, please know that it’s only possible with the generous support of our sponsor: Higher Logic.
On sharing data in an accessible format: “Honestly, basic charts are really powerful because they’re so well understood, they’re so well used, and it can be tempting to get into this, ‘I’ve created something nobody’s ever seen before, it’s so novel.’ … Generally, visualizations are for people other than just the person making them. Your audience is going to work that much harder to understand the message you’re trying to get across, so you’re creating some cognitive burden for them to just read it, let alone take away the point that you’re trying to get across.” –@tammylarmstrong
There’s no such thing as a free (software) lunch: “When you are giving up data, or when you’re getting free software, software companies aren’t doing this out of the goodness of their hearts. They’re getting something out of it, and most of the time it’s data. They’re monetizing that in some way, or maybe they’re latching on to your hard drive and leveraging your computing power for Bitcoin or something. There’s something that you’re giving up whether or not you know it, and we really need to be generally more savvy in terms of what that means.” –@tammylarmstrong
On making your data work for you: “Think about your goals with your community. Your data is a tool and a valuable asset, so if you start with your goal in mind and think about what you want to get out of the community, where do you want to be in five years, and then turn that around and say, ‘Well, where am I right now?’ Having that guiding point is super helpful because otherwise you can just get lost in all of your data. So, start with the end in mind.” –@tammylarmstrong
About Tammy Armstrong
Tammy Armstrong is passionate about using data to solve problems and make the world a better place. As team leader of customer service analytics at Wellmark Blue Cross and Blue Shield, she works to improve the customer experience and maximize efficiency. Previously, she worked in independent education fundraising as the director of analytics for Oregon Episcopal School in Portland, Oregon. Tammy also serves as the treasurer of Reboot Iowa, a nonprofit startup dedicated to educating adults in computer science.
In their spare time, Tammy and her husband Mike record Bright Lights Big Data, a podcast exploring how local communities work through interviews and discussions on urban planning and analytics. Tammy has an MS in Analytics from North Carolina State University, and a BA in Economics from Gettysburg College. She enjoys baking, playing soccer with her husband, and chasing after their toddler daughter and two cats.
- Sponsor: Higher Logic, the community platform for community managers
- Tammy Armstrong on Twitter
- Tammy on LinkedIn
- Wellmark Blue Cross and Blue Shield
- Oregon Episcopal School
- Reboot Iowa
- Bright Lights Big Data, a podcast exploring urban planning and analytics, recorded by Tammy and her husband, Mike
- KarateForums.com was the first community that Tammy joined and where she met Patrick
- The Children’s Online Privacy Protection Act
- Emily Temple-Wood, Wikipedian of the Year, on Community Signal
- Nexidia, an interactions analytics software
- Power BI, a data visualization tool from Microsoft
- Rise of the Evil Algorithm
- Fitness tracking app Strava gives away location of secret US army bases
- Ron Wyden and Ed Markey, two elected officials working to fairly govern the internet
- Threats to Section 230 Should Unleash the Political Power of Community Professionals
[00:00:05] Announcer: You’re listening to Community Signal, the podcast for online community professionals, sponsored by Higher Logic, the community platform for community managers. Tweet with @communitysignal as you listen. Here’s your Patrick, Patrick O’Keefe.
[00:28] Patrick O’Keefe: Hello, happy new year, and welcome to the first Community Signal of 2019. New year, new music for the show, as you probably noticed if you are a regular listener. Our guest this episode is Tammy Armstrong who works in customer service analytics at Wellmark Blue Cross and Blue Shield. We’re talking about speech analytics, big data, and where to start when you decide to analyze the mountain of data your community sits on.
Thank you to our supporting listeners on Patreon including Jules Standen, Rachel Medanic, and Serena Snoad. If you’d like to back this show, please visit communitysignal.com/innercircle.
Tammy Armstrong is passionate about using data to solve problems and make the world a better place. As team leader of customer service analytics at Wellmark Blue Cross and Blue Shield, she works to improve customer experience and maximize efficiency. Previously she worked in independent education fundraising as the director of analytics for Oregon Episcopal School in Portland, Oregon. Tammy also serves as the treasurer of Reboot Iowa, a non-profit startup dedicated to educating adults in computer science.
In their spare time, Tammy and her husband Mike record Bright Lights Big Data, a podcast exploring how local communities work through interviews and discussions on urban planning and analytics. Tammy has a master’s in analytics from North Carolina State University and bachelor’s in economics from Gettysburg College. She enjoys baking, playing soccer with her husband, and chasing after their toddler, daughter, and two cats. Tammy, welcome to the show.
[00:01:42] Tammy Armstrong: Thanks so much for having me.
[00:01:43] Patrick O’Keefe: I’ve know you for more than 18 years and we met because you joined a community that I managed, and still manage, KarateForums.com. You spent several years there, made thousands of posts, and became a moderator, and before the show you mentioned something that I hadn’t realized, which is that KarateForums.com was the first forum that you ever participated in. Do you remember joining?
[00:02:06] Tammy Armstrong: Yes. I don’t really how I found it, to be honest. Maybe my older brother found it first. I think I stuck with it longer than he did.
[00:02:14] Patrick O’Keefe: Sean?
[00:02:15] Tammy Armstrong: Yes, Sean, yes, good memory.
[00:02:16] Patrick O’Keefe: Spelled the right way, like my brother, Sean.
[00:02:16] Tammy Armstrong: [chuckles] Yes. That’s right, yes.
[00:02:22] Patrick O’Keefe: And my favorite, Sean Combs.
[00:02:24] Tammy Armstrong: Of course, of course. [laughs] I don’t remember a ton about the early days, probably made the obligatory introductory post and then was somewhat fascinated by getting my avatar belts [chuckles] updated, and ranking up as I posted more, and leading to the aforementioned thousands of posts, but yes, that was really my first interaction with forums and an online community. That was a formative experience, I’ll say.
[00:02:52] Patrick O’Keefe: You know, without dating ourselves too much, you were a teenager, I was a teenager. I’m trying to think, 18 years ago, yes.
[00:02:57] Tammy Armstrong: Yes. [laughs]
[00:02:59] Patrick O’Keefe: Coming at it as a teenager, and this is something we talk about on the show, the underrated contribution that teenagers make to the internet and this idea that that’s a community that has touched many people. I would say easily hundreds of thousands of people, possibly millions if you count people who just read a post and found value, and went about their lives, didn’t have to register. And yet, it was started by a teenager, a lot of my moderators have been teenagers. Of course, always respecting COPPA, the Children’s Online Privacy Protection Act, but I was always impressed with the maturity that you brought to the community as a teenager.
I don’t know. It’s just one of those things that if you talk to people who work in community who I’ve work with volunteer programs, more often than not they will say the same thing, which is that a lot the communities that are well managed that are not online major corporations, that are volunteer-based, are often staffed by teenagers who are doing a great job behind the scenes to maintain the value of this thing that many adults derive value from, but don’t necessarily think of it that way. I had a guest on the show, Emily Temple-Wood who was Wikipedian of the Year a few years back.
She made a comment that if I ran up to someone on the street and held up my written article about this woman scientist, and I am a teenager, they would look at me [chuckles] like I’m crazy or suspiciously. They would not value that information the same way that they do when they go to Wikipedia, open the article, and don’t see the face, or don’t know who wrote it. When I think of your time at KarateForums.com, I think of that sort of thing, of the community that you helped to moderate and create, and then keep going on the right track. I’m glad that we’re still in touch after all these years, but also it’s just a great reminder of the value that teenagers give to the internet.
[00:04:39] Tammy Armstrong: Well, of course, you know better than most people that home schooled teenagers are especially special. [laughs]
[00:04:43] Patrick O’Keefe: Right, yes.
[00:04:46] Tammy Armstrong: I think it’s easy to get older and forget what it’s like to be younger, right, and to have that passion and that focus for something that you can bring if you’re really excited, as well as frankly the free time. The older we get, the more responsibilities we accumulate, and it’s harder to dedicate that much passion and focus to one thing. You start having a paying job, and maybe kids, and just other responsibilities. That’s not to underrate the other qualities that teenagers can bring to those efforts. I think you alluded to kind of a maturity. Again, I think people expect way less than is possible.
You and I both dealt with this on the forums. I think people were surprised to find out you were a teenager at the time, running these forums. Then, if they did find out, would sometimes try to use that against you in some way that they disagreed with how the community was being run. Usually because they’re running afoul of the rules. [laughs]
[00:05:41] Patrick O’Keefe: Right, right. I had to exercise that authority and power trip from my mom’s basement. That’s just what it was, right? Exactly.
[00:05:49] Tammy Armstrong: Yes, but to your point, I think for many years people have recognized the internet as an equalizing force, right? Everybody can put their ideas out there, especially at the time period that we were out there starting these communities. There wasn’t Instagram and Facebook and these ready-made platforms. It was something that you made from scratch and it was something that the community members built along with it.
You were literally creating something out of nothing as a community in a way that we probably take for granted more now because we have these platforms set up in broader social networks. The enthusiasm and dedication and creative thinking, you know, teenagers don’t really have those boundaries that they think about in terms of what’s possible. I think they can be a lot more optimistic. I think that can really come together in a safe space in a beautiful way.
[00:06:40] Patrick O’Keefe: It’s not uncommon these days to interact with someone that you mentioned, that is a local person, local celebrity, serious celebrity, whatever. One of the funniest things I remember from your time at KarateForums.com is that you mentioned someone-
[00:06:56] Patrick O’Keefe: -and I don’t even remember what she did. She was like a fitness person or maybe a kick boxer or a martial artist. I don’t need to mention any names. Actually, I do remember her name, but it doesn’t matter. You mentioned her in a post and I think she was small enough where she was Googling herself and found the forum, and came in and was like, “What are you talking about here?” I brought her back to reality, I remember. That was one of the funniest things ––
[00:07:18] Tammy Armstrong: Oh my gosh, I was so embarrassed when that happened. [chuckles]
[00:07:20] Patrick O’Keefe: -that I remember from that time, I don’t even think– You didn’t even say anything all that bad. I have the record somewhere.
[00:07:25] Tammy Armstrong: I remember it very clearly. [laughs]
[00:07:28] Patrick O’Keefe: Yes, that was really funny.
[00:07:30] Tammy Armstrong: Yes. I’m embarrassed about it, but it was also like a really good eye-opening and learning experience for me that even though I thought of Karate Forums as being an insular community, it was still open to the public. Somebody could Google themselves and find that I had mentioned their appearance in a magazine and I didn’t like the fact that they were wearing makeup, which, you know, my feminism has greatly-
[00:07:53] Tammy Armstrong: matured since then. I didn’t think that those two things could be consistent at the time, and just realizing “Yes, I should probably watch what I say and be a little more respectful of other people’s feelings.” [laughs]
[00:08:06] Patrick O’Keefe: I want to move on from that past that we have together, but one last question. Is there anything that you take away from your time spent as a moderator that you think about, or apply to the work that you do today, or that sticks with you in any way?
[00:08:20] Tammy Armstrong: I think a really broad lesson is just how people can grow. I was recalling my time on Karate Forums and remembering that I myself, before I was a moderator, got a private message from a moderator saying, “Hey, you might want to think about this post that you just had,” and giving people the benefit of the doubt that maybe– Is it Oprah that says, “When we know better, we do better”? [laughs]
[00:08:42] Patrick O’Keefe: Yes.
[00:08:44] Tammy Armstrong: Sometimes we need to know better so we can do better and trying to give people that benefit of the doubt before jumping to conclusions.
[00:08:51] Patrick O’Keefe: Let’s take a moment to talk about our great sponsor, Higher Logic.
Higher Logic is the community platform for community managers. With over 25 million engaged users in more than 200,000 communities, organizations worldwide use Higher Logic to bring like-minded people together by giving their community a home where they can meet, share ideas, and stay connected.
The platform’s granular permissions and powerful tools, including automated workflows and consolidated email digests, empower users to create their own interest-based communities, schedule and manage events, and participate in volunteer and mentoring programs. Tap into the power your community can generate for you. Higher Logic, all together.
You work with customer service analytics. What are you measuring right now that you find to be very interesting or meaningful?
[00:09:34] Tammy Armstrong: Yes, there’s so much. I think right now one of the challenges that a lot of people have, and you mentioned this earlier in terms of how much data forums and everybody really is sitting on and how are they using it, in the healthcare space and lots of industries, it’s how do we use all of that data in a meaningful way, how do we get the most out of it. One thing that we’re really trying to do is to look across all of our different channels and understand not just, “Okay, this is what happens on the phone calls, and this is what happens on the self-service channels and the emails,” but really marry them and follow a single person and have a more customer-centric view.
Understand they tried email first, and then they called, and why did they do that, and how can we use all of those insights to make the experience better? That’s a squishy concept that involves a lot of things, and we’re still honestly trying to figure out how best to measure that, but it combines a lot of structured and unstructured data in terms of not just the events themselves and where they happened and when, but also what was the content of them. What was the text of the email, or we even do speech analytics, what was contained in the phone call itself. It’s a really big, messy problem, which is part of what makes it so exciting.
[00:10:49] Patrick O’Keefe: You touched on something that gets talked about in community a lot, call deflection and the idea that if someone receives their answer through a channel that is more cost effective, that you saved money. The biggest problem with call deflection is that if you don’t have unified data you don’t, in my opinion, actually have call deflection, because as you alluded to, you don’t know if someone tweeted, and then emailed, and then called, and then visited. [chuckles] You don’t actually know if they channel shopped before they actually came to their answer.
In community, we look at that from the sense of people getting a question answered by the community. That could be in a number of different forums, forum comments. If someone comes to the forum, they ask a question, they get help from a community member, they didn’t have to take up any of our time. Now, that is how a lot of people choose to look at it and maybe calculate some sort of value per answer. But if you don’t know or have that account tied to the member’s email tickets, or their live chat logs, or their account through their phone calls, you really don’t know what’s going on there, and you could actually be not saving any money.
There might not be any savings at all, there might not be any deflection, they may have still called. That sounds like something you’re thinking about. What are you finding interesting, or helpful, or useful, or what would you think is the first step of bringing that data together?
[00:12:11] Tammy Armstrong: Well, the technology behind it is the biggest barrier, I would say. I think a lot of contact centers have really disparate solutions. They have one vendor that they use for capturing their phone interactions, which honestly there’s like four or five components that go into just getting someone on the call. You’ve got the telephone prompts that they answered, maybe by voice or by touch, to get to a representative. Then you’ve got the call recording, then depending on if they’re reaching a customer service rep, there’s probably some kind of CRM tool that has a case on it.
That’s already three systems for one channel [chuckles] to marry together and make sure you can map across all those different data sources and have common identifiers, and hope that they didn’t go wrong somewhere. You now think about maybe you’ve got a larger customer database with sales information in it that may not be tied directly to your customer service case database. Well, you want to understand that, right? Did they call two days after their package was delivered, therefore, talking retail?
Then they’ll have just a completely different vendor solution for something like chat or email. Those tend to be created in a very siloed way if you’re not intentional about it up front, and so have common identifiers, like you said, is really a big challenge to start with. You can make some assumption sometimes, but a lot of times it ends up falling on us in analytics to do some band-aid solutions to say, “Well, I know that this identifier maps with this. No, system A maps to system B. I don’t even care what’s in system B, but I need to use it to map to system C.” You put these behemoths together.
That’s just table stakes. Can you even look at what all is out there? Then, I think one of the surprising things when it comes to channel deflection conversations is that you almost need to bring a design thinking approach to things. Why are people using one channel versus another? I go to conferences and you’d be surprised the number of people who will say, “Well, we just decided to–” When people of this certain customer type call, they’re– Or maybe it’s a–
I’m trying not to get very specific about what I actually do, because I’m not strictly speaking allowed to talk about it [chuckles] very much. Maybe not a customer but a supplier or somebody that calls into a customer service line and you just say, “You know what, you’re not allowed to call about this anymore. You have to self-serve or you have to use the web system.” That doesn’t work.
[00:14:37] Patrick O’Keefe: Right, no.
[00:14:38] Tammy Armstrong: You run into them the next year and they say, “Yes, we stopped doing that.” You can’t just shut off a channel altogether and expect it to fix the problem. There’s a reason that people are going to one channel versus another. Maybe it feels easier to them, maybe they feel like a phone call is being recorded and they have a reference number that is proof that they checked with you in a way that they don’t have when they go onto a website that could change tomorrow.
There’s a lot to be thinking, and putting yourself in your customers shoes and designing around that. Those are the same concepts that Apple used to create a great mouse, a computer mouse that people loved and other products that people loved. I think people don’t always value those skills as much as they should in a field like customer service and even in analytics in general.
[00:15:24] Patrick O’Keefe: Yes. These are very relatable problems for community pros, because we do have a disparate set of tools. I have a disparate set of tools at my day job. This is the constant joke complaint venting point for any conversations that relate to measuring the value of what we do, because I want to show that people who are active in our online community give us money at a higher rate than people who do not. I want to know what that percentage is, how much more is that customer worth to us. Then, when we start to see that picture, we start to unlock the value of what we’re doing.
When you talk about those disparate systems, and I hear you walk through it, to me, one thing that jumps out is, unfortunately, if you really want to unify your data, it’s not going to be something that the vendor is going to typically step up to do. It has to be the type of thing that you have internal tech resources behind and budget to say, okay, we can write something, we can hook into there APIs, we can pull it into a dashboard and unify this data to the greatest of our ability, but it has to be us. It’s not going to be them.
[00:16:32] Tammy Armstrong: Yes, or at the very least making sure that you were advocating for that at the time when the vendor solution is implemented, because it’s probably going to incur some additional cost in time and you have to be willing to make that investment, or you’ll regret it down the road. [chuckles]
[00:16:48] Patrick O’Keefe: I’d love to hear a bit about speech analytics and speech analytics software, and how you use that to find new opportunities and to improve the experience for the user member and the person on the other line. Talk about that a bit.
[00:17:03] Tammy Armstrong: Yes. There’s a couple different kinds of speech analytics technologies out there. We use one called Nexidia, which was founded by– I want to say a University of Georgia professor who got it patented and started this company, where he basically uses a neural network to transcribe speech, but not from audio to words, but from audio to phonemes. Phonemes are the smallest parts of human speech. I believe there are about 50-some phonemes in the English language, and these basically represent sounds.
The beauty of doing that transcription is that you can find non-dictionary words in your media. Instead of limiting yourself to a one-time transcription in the healthcare space, we might be looking for calls where customers are mentioning claims, and if the transcription service accidentally transcribes that to clams, I’m never going to find claims [chuckles] in that call. Instead, I can teach the system what that word sounds like phonetically, and then I can find it, and it’s actually really good at handling things like different kinds of accents.
There’s two approaches to, I would say analytics in general, and this applies to speech analytics particularly well, which is having an unsupervised approach versus a supervised approach. That’s a machine learning distinction. An unsupervised learning approach is where you basically say, “I don’t have a goal in mind. Just show me what you can find.” That might be where we don’t say, “Show me all the calls where they mention claims,” but just say, “Go ahead and try to transcribe words, and show me what the most common ones are. Show me which ones are more common when calls are longer, or when customers sound angrier.” Some of these softwares have proprietary models to measure sentiment.
A supervised learning approach would be where you’re saying, “I know that I want to find my customers who are using the “Contact Us” page of the website. I want to start finding very specific phrases so that I can quantify how often they say on the lower right-hand side of the screen, so that I can start to triage. The nice thing here, too, is that you can measure before and after, maybe you make an update to your website, and then you can see, “You know what? We were getting a hundred calls a day about this before we made the change. Now we’re down to two without doing any additional work,” because you’ve trained the system how to find those conversations.
[00:19:34] Patrick O’Keefe: It’s really interesting. Obviously, with online communities it’s not holy texts, but it’s often texts because it’s easier to be present in a conversation that exists over a period of time, it’s asynchronous, where you don’t have to be present in the moment, you also don’t have to spend an hour watching someone say why TKD is ineffective in the street. You can just have someone write their post and then read it, and go on with your life.
It’s interesting to me because there’s so much text and there’s so much information, and there are so many different ways to analyze that data to find new information. For me, where the most benefit often comes is solutions that are accessible to people who don’t have a lot of money. I’m sure it’s obviously the case for big data, and for analytics, and different services. You can spend a ton of money real fast. There are very expensive community software platforms and vendors that serve this industry, but when people say, “We did this great thing in this software that costs a hundred grand,” I’m like, “That’s cool, but that’s not going to impact anything.”
Because I don’t care, I don’t know, what Nabisco has for their community feature. That’s not going to make its way down here. My thing is when that feature gets put into a platform that is a few hundred bucks, or better yet, is open source. Then we can actually impact online communities for the better. Whether that is filtering tech that used to be more expensive and is now more affordable, or people can more accurately filter content and knock down the things that machine learning and AI does well while not throwing out the baby with the bathwater, so to speak, and just taking the more nuanced cases and thinking that the algorithm can read through them.
It’s such an interesting thing. It’s one of the reasons that I advocate for mature community teams to have a data person. I think it’s super relevant. We have so much data, so much information. Unfortunately, a lot of community pros, it falls to them. The thing is, I have a three-person team that I manage now. Four if you want to include me in there and look even better. So, four. A team of four. The reality, though, is that a lot of people have one person and that’s the team. That’s all everyone ever sees, and that’s why people leave this work and do something else.
In those cases, they’re often tasked with also being the people who analyze the data and try to come up with insights and solutions. This is a long-winded way to make it to my next question, which is– I had first thought I wanted to hear about your favorite data visualization tools. I want to hear about that, but after your favorite data visualization tools, if there are any tools that you think would be useful to community pros who have databases of content, usually SQL databases but could be other types of databases as well, of content, tools that you think might be useful in those cases. Just put some data in and see what happens. I’d love to hear what tools you find meaningful.
[00:22:30] Tammy Armstrong: Yes, absolutely. I consider myself an amateur data visualization geek. I love data visualization, so I love that question. This may be an unpopular answer, but honestly, Excel can be great for visualization. I think it’s underrated. You just have to know how to massage it and get rid of all of the defaults, because they’re usually horrible.
[00:22:53] Tammy Armstrong: Gridlines and outlines, and for God’s sake, don’t use anything there-dimensional. [laughs] If you can get around some of the dark paths it can try to send you down, Excel actually offers you a great amount of control, some really great powerful basic charts. Honestly, basic charts are really powerful because they’re so well understood, they’re so well used and it can be tempting to get into this, “I’ve created something nobody’s ever seen before, it’s so novel.” That’s a fun exercise, but it also means that your audience– Generally, visualizations are for people other than just the person making them. Your audience is going to work that much harder to understand the message you’re trying to get across, so you’re creating some cognitive burden for them to just read it, let alone take away the point that you’re trying to get across.
I really love Excel. There are some great blogs and podcasts and books out there that actually do focus on how you can do everything that they promote in Excel and other commonly used tools. That said, if you want something a little bit fancier, Microsoft. Again another tool from Microsoft, but Power BI is something really great. Something that we’re working with more lately at my organization. They actually have a totally free version that you can download. I’m pretty sure this is true, I think I have it on my personal laptop. You can download Power BI, connect it to hundreds of data sources.
[00:24:17] Patrick O’Keefe: You quoted the page. I pulled it up right here while you’re talking. It says literally, “Connect to hundreds of data sources,” so you nailed it.
[00:24:21] Tammy Armstrong: [laughs] Nice. I will be collecting my check from Microsoft later on. Whether that’s SQL, or flat files, or Hadoop or just anything you can think of, probably Salesforce all these different things. I think the only time you really need to pay for Power BI is if you want to be able to publish things to a cloud platform for an enterprise. They have really democratized open source storefront to download types of visualization that other users have created, which really lets you get outside the box of what Excel would give you.
Also, if you have somebody else, honestly, you could just email somebody a Power BI file like you would email them an Excel file. If they have Power BI themselves, they’re going to be able to interact with it, and filter it, and do things like that.
[00:25:08] Patrick O’Keefe: Very cool.
[00:25:09] Tammy Armstrong: I also have to, of course, give a shoutout to programming languages like R, for those who really want to get their hands dirty, but there’s a steep learning curve to something like that.
[00:25:18] Patrick O’Keefe: Moving off of hardcore data side of this, what are your big concerns? I didn’t mean to use big concerns, but I’ll stick with it. Big concerns about big data. You wrote, I think it was a while back, about the rise of evil algorithms on your LinkedIn page, and you have a young child as well, so I’m sure that factors into this conversation. What keeps you up at night? What really concerns you about the work that you do and the power that you probably see that it has?
[00:25:45] Tammy Armstrong: [chuckles] Yes. I think it’s really tough. There’s so much of it that’s probably going to be unavoidable. A lot of it is just how much of our data is out there and all the new ways that people are thinking up to use it. This is something my husband and I were talking about recently. I think for a long time people have thought of the basics, your social security number and your bank account information as being the stuff you really needed to protect in terms of your personal data, but as more and more of our behavioral and activity data is being shared in ways that we might not think about, that app that I’ve added to my phone, or that I’ve authorized on Facebook that’s collecting stuff I didn’t really think about.
As people start putting that together in different ways, you think about Cambridge Analytica, you think about Strava, the beginning of 2018 and the map that they released that inadvertently highlighted secret military bases. [chuckles] There’s so much that we don’t know to be careful about. It’s hard sometimes not to want to put your little tinfoil hat on in those scenarios. When it comes to my daughter, I think I’m nervous for the day when she’s a tween or a teenager, using technology that I’m totally unfamiliar with, unaware of the dangers of– Just because I’m already like, “I can’t use Snapchat, I’m too old.” I tried and I never use it. [laughs] What is that going to be when she’s older?
At the same time, back to our earlier conversation about how teenagers can lead the way, I think just trying to stay as educated as possible, put that critical thinking skill set to use and try to support her as she navigates these changes. I think that’s the best we can do. How did our parents feel about us when the internet started and chat rooms were a thing, and community forums, and we had to say, “No, you can’t just go meet somebody you met online. You’ve got to think about how you share that information.” This may be just the next frontier of that.
[00:27:48] Patrick O’Keefe: You didn’t meet your husband online, did you?
[00:27:49] Tammy Armstrong: No, I met him in Costa Rica.
[00:27:51] Tammy Armstrong: Even better. [laughs] Honestly, I think that freaks him out a little more than if I had met him online. [laughs]
[00:28:00] Patrick O’Keefe: At the start of the conversation we talked about KarateForums.com and how you mentioned the idea of having to create a community yourself more so than just having something that’s ready-made and sort of a social network, I guess, in a lot of ways. When I look at a lot of the ills that I would maybe tie to the collection of data, assigning blame is not what I want to do, but I think that a lot of things have grown worse as we have made fewer choices and given a few platforms the overwhelming lion share of our data.
One of the things that concerns me is when people talk about legislating the internet and legislating speech online, I’m not wholly against legislation, that’s not the point, but what happens is that the same legislation that applies to this will almost certainly in one way or another apply to the collection of data, how the data is housed, what you can have, how long you can have it, when you have to remove it, what you can use it for, et cetera. These are very parallel conversations when you talk about the law. The legislation is written for Facebook. [laughs] I think that’s really the point. That’s what I’m getting towards. The legislation on this that people are threatening if your representative there in Iowa ––
[00:29:16] Tammy Armstrong: Steve King?
[00:29:17] Patrick O’Keefe: Yes.
[00:29:18] Patrick O’Keefe: Steve King, up there talking to the Google CEO how his grandchild picked up a phone and he saw nasty things in an ad about her grandfather, and he mentioned, of course, that the phone was a hand-me-down, meaning someone else used it before that child, and that someone probably may have searched for Steve King, which led to the ad being served, of course. The reality is that these are our lawmakers. They’re going to write the legislation for Facebook.
They’re not going to write legislation for KarateForums.com, and yet the same legislation that will apply to a multibillion-dollar corporation, if we’re not careful, will end up applying to a one-person community that doesn’t even make enough to pay one person [chuckles] to be on salary. Section 230 is the basis for online community moderation in the U.S., the Act that gives us the right to moderate without being liable for what we leave up. I don’t know if it was him who brought it up or someone else who was also ill-informed on that same panel, but brought up Section 230 as a thing that dangled around and maybe Section 230 needs to change if the search results don’t look better for conservatives.
[00:30:26] Patrick O’Keefe: There’s a lot of unintentional humor here. I’m triggering a lot of laughter in you, which I appreciate, but also if Section 230 is meaningfully impacted, that’s going to impact my communities and impact the work that I do. It’s interesting because we didn’t necessarily have that. When we were talking about KarateForums.com, you joined in 2002. The Section 230 of the Communications Decency Act was passed in ’96. The COPPA, the Children’s Online Privacy Protection Act was passed in ’98. These are– Especially the CDA, CDA is a big deal.
The internet was very different then. We have a different, more spread out environment of what the internet was considered. Now, all the data, it feels like– Not all the data, but most of the data, it feels like, falls under Facebook. I think that scares people in a lot of ways, and now we’re going to have to deal with the consequences of that. I’m not saying we haven’t all been willing participants in one way or another, although certainly Facebook has been programmed in a way to keep people there more than anything else.
Gosh, where was I going with this? I think the thing is, though, when we look at how we met, KarateForums.com, and we look at how the internet was, and we look at how it is now, and we think about the concerns of big data and the concerns of how our information are going to be used, I think it’s really relevant what the current state of the social web is, which is there’s a lot of eggs in one basket.
[00:31:55] Tammy Armstrong: Yes, I think it’s the next wave of computer literacy. It’s not just how you use the mouse and the keyboard and all of those things, it’s how do you understand how these things work. I think a lot of people don’t think about the fact that when you are giving up data, or when you’re getting free software, software companies aren’t doing this out of the goodness of their hearts. They’re getting something out of it, and most of the time it’s data. They’re monetizing that in some way, or maybe they’re latching on to your hard drive and leveraging your computing power for Bitcoin or something, but there’s something that you’re giving up whether or not you know it, and we really need to be generally more savvy in terms of what that means.
You alluded to legislators legislating on things they don’t necessarily understand, and that’s definitely relevant in the analytics community as well. I don’t know that we’re dealing with a lot of upcoming regulations for analytics as an industry, although finance certainly deals with that that you have regulators who come in and are going to look at your models, so you basically can’t make them too complicated, because otherwise they won’t understand what they’re doing so they can’t regulate them.
We do see it with things like statistics being used in gerrymandering cases and the models being too complicated for the judges to rule on. They can’t understand what’s being put in front of them when people are trying to come up with measures of gerrymandering and whether voting districts are unequally designed or designed with, say, racism in mind, or something like that. There’s this lack of understanding of the methods that makes it really difficult to do anything with the substance of what comes out of them.
There’s a part of me that’s concerned that what’s happening with Facebook and the legislation that’s coming out of that, and even things like GDPR in the EU, that my industry might see something more like that in the future. That all it’s going to take is one or two well-publicized gaffes for somebody to want to start regulating and saying, “Well, if you’re going to use analytics or statistics in a company of this size, then you’re going to have to have regulators look at all your models.” What a nightmare [chuckles] that would be, trying to get that done. It’s hard when you’re dealing with something very technical that has large impacts. How do you actually get the legislators to understand that sufficiently to legislate well?
[00:34:21] Patrick O’Keefe: As I was leading to Steve King, I actually forgot you were in Iowa, and so that was like a happy meeting in the middle there of two thoughts that came together. I thought, “Oh, I arrived at Steve King. Oh, Tammy, Iowa, Steve King, association.” [chuckles] That’s why I associated it with you, of course. To paint this picture is not to also say that there aren’t people who get it, right?
[00:34:42] Tammy Armstrong: Right, right.
[00:34:43] Patrick O’Keefe: Because there are people who get it. In fact, one of the co-authors of Section 230, Ron Wyden, is still [chuckles] a senator now. One of the things that frightens me, frankly, is that he’s recently raised the alarm that Section 230 might be in danger. That’s a concerning thing, but there are people who get it. I had Senator Ed Markey from Massachusetts came out to my apartment building here in LA and spoke to some of the residents, because we have people in the building that have a lot of followers on different social platforms. It was part of the net neutrality effort that ultimately failed.
There are senators who get it, there are senators who have been fighting this fight for a long time, representatives who have done the same. It’s one of those things where I don’t want to repeat myself too much, but on the last episode of this show, we talked about the political power of community pros. There are as many of us as there are coal miners, and the reality is that we do have platforms that have a lot of reach, and we just don’t use those platforms to spread political messages because it, in general, would not be appropriate.
If there is legislation that threatens those online communities and the ability to manage them appropriately, then I think in some ways it’ll be time to activate, to say, “Okay, well, if you like this community, then you might want to call your representative.” [chuckles] Because if this law changes, then maybe I don’t want to deal with that when it comes up, because I don’t have the time to manage that situation. GDPR brought that up. A lot of fear around GDPR for a lot of people who don’t have two nickels to spend on anything to do with GDPR for their online community.
Because we forget that the vast majority of online communities and social spaces are still run by people who like that thing. [chuckles] We have a few massive corporations, and sure, there are online communities that have become businesses with full-time staff, and that’s great, but most of them are just people who just like the thing and wanted to talk about that thing, and so they started a community, and now they have to learn what GDPR means [chuckles] and try as they might to satisfy that. Many people rather say, “You know what, if something happens, I’ll just shut the door.” That’s unfortunate.
[00:36:48] Tammy Armstrong: That could even create an opening in the market for “free community software” that lets your smaller in-house forums like that be in GDPR compliance, be in compliance with those things that they don’t want to have to think about. Suddenly, they’re not in control of it anymore, and there’s probably some fine print that they’re harvesting all of this data, too. As these things become more complex, it creates an opening for somebody else to take advantage.
[00:37:17] Patrick O’Keefe: Yes, for sure it does. To wrap up that conversation, talking about your daughter, I guess the reality is that by the time she gets interested, my feeling is that we’ll be in a very different place. There’s a pendulum swing almost that goes back and forth between identity and real ID, sharing information. My brother who’s 13 years younger than me, he doesn’t have a Facebook account and he doesn’t want one, and I don’t think he’ll ever have one. That’s not even that big of a gap there. The reality is that there are a lot of kids who don’t want a Facebook account, don’t have one. That’s just not what is in their vision.
[00:37:55] Tammy Armstrong: It’s bad for their mental health at this point, honestly, yes.
[00:37:58] Patrick O’Keefe: Yes, it is, but I’m just saying as a matter of choice. They don’t use it, they don’t want to. It’s not cool, it’s not interesting. We are going in a different direction of maybe smaller players that’ll become bigger players, but maybe not as big as this. I don’t know. We’ll see. No one can predict the future. [chuckles]
[00:38:17] Tammy Armstrong: Yes. I hope in my heart of hearts that she is part of a generation that just eschews technology and just wants to hike all the time. [laughs]
[00:38:26] Patrick O’Keefe: Goes back to being a monkey girl and doesn’t have to out themselves as– [chuckles] I have that name. Yes. Last thing I wanted to talk about is kind of a pie in the sky open-ended thought, as I guess many of mine are. As we talked about before the show, online communities sit on just incredible amounts of data. Even these independent communities like KarateForums.com that’s going to turn 18 in May. 18 years of content, right? 18 years of member data, posts, and of course, major corporations have massive amounts of data in their online communities, too, but we often host our own databases, right? We have full access to that data.
I’ll phrase this two ways to make it easier to approach. First way is, looking at it from the outside, where should we start? Or if you prefer, if you were going to take a role in an online community that was mature, well-established and just had a database of content that they hadn’t really had the opportunity to really start analyzing or think about in any meaningful way, where would you start?
[00:39:30] Tammy Armstrong: Well, there’s so many good opportunities there. One of the things I’d be most interested in, I think, is basically doing back to that idea of unsupervised learning, some clustering techniques. That’s a technique that’s often used in marketing as well as a lot of other industries and applications. Basically, toss all of your data that you’ve got at an [chuckles] algorithm and say, “You know what, send me back five groups. I want to know what are my five kinds of users.”
That might come back with, “Okay, you’ve got your male teenagers who practice TKD and spend a lot of money on online gaming or something. Then you’ve got your stay at home moms in their 30s,” and you come up with these different almost personas and just see what that’s like and maybe map that against different levels of activity. Something you mentioned earlier in a slightly different tack was talking about active users and how much money do they spend. It got me thinking about how do you define an active user? That could be a subjective term.
[00:40:34] Patrick O’Keefe: Yes, it is.
[00:40:36] Tammy Armstrong: [chuckles] Now I’m trying to figure out where those breaks are. What I generally advise is, unless you’re just playing to play and see what comes out, is just to really think about what are your goals with your community. Your data is a tool and a valuable asset, so if you start with your goal in mind and think about what do you want to get out of the community, where do you want to be in five years, and then turn that around and say, “Well, where am I right now?” Having that guiding point is super helpful [chuckles] because otherwise you can just get lost and all of your data. So, start with the end in mind.
[00:41:10] Patrick O’Keefe: Tammy, it has been a pleasure to have you on the show. I’ve enjoyed the conversation. Thank you for making time for us.
[00:41:15] Tammy Armstrong: Thank you so much for having me. It’s always a pleasure to catch up.
[00:41:19] Patrick O’Keefe: We’ve been talking with Tammy Armstrong, team leader, customer service analytics for Wellmark Blue Cross and Blue Shield, host of the Big Lights Big Data podcast which explores how local communities work through urban planning and analytics. Visit blbdpod.com.
For the transcript from this episode plus highlights and links that we’ve mentioned, please visit communitysignal.com. Community Signal is produced by Karn Broad and Carol Benovic-Bradley is our editorial lead. Until next time.
If you have any thoughts on this episode that you’d like to share, please leave me a comment, send me an email or a tweet. If you enjoy the show, we would be so grateful if you spread the word and supported Community Signal on Patreon.