Name Testing Gold Standard

The NameStormers team is dedicated to their creative process, but no team member invokes pride of ownership, instead, we always aim to turn to the data. Since the foundation of NameStormers was built by four former employees of Nielsen Media Research, gathering quantitative data is believed to be crucial to our creative process. From developing one of the first AI platforms back in 1985 to dabbling in leveraging modern-day AI, listen to this week’s episode to learn more about how Mike’s team structures their research process and how we optimize results. 

Episode Seven of “Naming in an AI Age” Podcast 

Adelaide Brown: 

Hi, and welcome back to “Naming in an AI Age” with your NameStormers team. Mike Carr and I are back again, ready to talk a little bit more about the analytical side of things, um, for the name testing “Gold Standard.” Um, so this conversation has been inspired by some of our past clients, um, those who have hired us for the name, um, creation and development and that kind of consulting process. But then also then to go into, um, name testing, which we’re just gonna talk a little bit more about what that means. We’re gonna jump right into it, if that’s okay with you, Mike.  

Mike Carr: 

You bet.  

Adelaide Brown: 

So, when thinking about name testing, and we’ll go into in future episodes, kind of the details between the type one, type two, but how many names would you say is the ideal number when conducting market research testing?  

Mike Carr: 

You know, that’s a really challenging question to answer because so many clients come to us with different practices in place. But what we would recommend, and, and I’ll share with you why and what hasn’t worked for us, is you don’t really need to start with more than a dozen. And so, clients will say, well, we haven’t, whittled the list down, you know, we have 25 or 30 or 40 or 50, and we want to use some kind of a conjoint or max diff approach to let our customers whittle those names down. And we don’t think that’s really the best way to do it. First of all, most of those names or many of those names won’t be available from a legal perspective. So, what we will often do is pre-screen the names for trademark hits, web hits, .com usage, you know, common law usage. Maybe we’ll knock out a huge chunk of those.  

It’s usually over half. And then there’s always a risk associated with what’s left, right? That every name then has, you know, some type of a- think about a stoplight- green, yellow, red risk assessment. And so, the test, you don’t wanna test all high-risk names or all low-risk names. And so, we typically would recommend, well let’s pick some names that look like you’ve got a pretty good shot at getting through any kind of legal hurdles, registration hurdles, regulatory hurdles, and then maybe look at some names, a little bit higher risk, and then you get down to maybe, um, six to eight names is pretty ideal. The way we do testing the way a lot of other folks do testing. It allows you to get I think, more real-world results back from the target. Some folks still feel that a magnetic approach is the only way to go, which is, you only present a single name to a single respondent. We don’t prefer that for a variety of reasons. Um, but we do think if you try to test too many names, it just sort of negates the validity of the test overall cause there’s just too much confusion and there’s too much bias created by the time they see that 13th, 14th name. So, the maximum for us would be 12. The a minimum would be three or four, and the sweet spot would be six to eight. 

Adelaide Brown: 

That’s so great that y’all also kind of guide that path, not making sure or making sure that there aren’t too many high-risk names. Finding that right balance amongst those three to four or 12 names. Um, can you walk me through the names from your research process? So, if a client were to enter a contract for that market research, how, what are your first couple steps and what are the key process points?  

Mike Carr: 

I think probably one of the, the biggest things anybody that’s doing this to think about is how do you present the name in as real world a context as possible? We’ve seen some folks that like to do a focus group and they’ll put each name up on a two foot by three-foot whiteboard all in the same font and all in black without any context. And so, their first exercise is people sit around a table and you throw up a name and you say, what does this name remind you of? Or what do you associate with this name? And I quite frankly just think that’s a total waste of time because no one ever sees a name without some context wrapped around it. So, the most important thing is to establish what’s that context going to be. So, like if it’s a product that’s gonna be in a brick-and-mortar retail store, can you show it on a dummied-up package?  

Right? How would it actually look on the box? Um, if it’s a digital offering, can you portray it digitally, whether it’s on the homepage or the, the first screen you see on your app or whatever that might be, hopefully with some, some graphic elements, right? Maybe your color palette, maybe your picture, uh, an ad mockup. It doesn’t have to be perfect, but whatever you can do to sort of help it bring alive for the respondent so that when they give you their feedback, they’re giving it to you based upon the context in which they actually are gonna see the name or hear the name or talk about the name versus something that’s very artificial and very abstract. So that’s, that’s like the foundation of the research. If, if we can’t do that, I’m not sure all the other things we’re trying to do are really gonna be worth, worth it. I can keep going, but if you have specific questions that you want to ask me, fire away.  

Adelaide Brown: 

Yeah. I know you’ve talked about there are several companies that’ll do the system one, but we in particular do the system one/system two. How did y’all decide to implement that specific research technique?  

Mike Carr: 

Right, so this comes from Daniel Kahneman’s book of, of years ago, you know, “Thinking Fast, Thinking Slow,” and you know, I think he won a Nobel Prize for that book, or at least some of the research that he’s done. And, and basically to me, the way I think about it is, you know, we, we tend to be lazy people and our brain, our brain especially likes to do things the easy way. So, and it’s sort of built into our brain and our evolution, right? That it wants to free up as much of the brain’s horsepower for the things that are really, really important. And the brain uses a disproportional amount of the energy the body consumes. So, the less juice it uses, the better cause that keeps energy for all kinds of other things. So, this idea that, look, we’re lazy creatures in anything that can be dealt at the subconscious versus the sub, the conscious levels the default.  

And that’s what System one thinking’s all about. That for many things we don’t consciously think about them, we simply react to them. And that’s the way naming works. Very few people that we’ve ever talked to consciously think about, well, do I like this name or not? They don’t, they just, they see a name and they either react to it positively or negatively. And if it grabs them, they may pick up the box or they may pay a little bit more attention to the ad and may click through the website. And if it doesn’t grab them, they’re gone and it’s like a second or two. And so, what we do in our research is we measure that. We, we look at how quickly and what order and some other things, so we can get an assessment on is this name working at that subconscious level, which is the most important level.  

What most people do, uh, is they ask questions like, well, do you like this name? Does this name fit the positioning? And those aren’t terrible questions to ask, but you have to understand that 99.9% of your customers will never ask those questions. And as soon as you ask those questions, you then move into the system two thinking, which is that more rational, deeper pondering. And this is somewhat dependent on the price of the item, right? If you’re, if you’re buying a $10 or $20 sort of impulse purchase at the grocery store, yeah, not a lot of thought goes into that. If you’re buying a, an automobile or you’re buying a $5,000 mattress, it’s gonna be a more considered purchase. And so that’s when I think those more system two questions are really valid, right? That you still want that system one, you still want the name to quickly grab you, grab the target so that you can then tell the story, or you can bring them in and have them further investigate the product. But then, you know, they are gonna be thinking more about the product, its attributes, its benefits, and perhaps even the name. And so that’s why both system one and system two are valid, but the weight you give each one varies based upon the product, the price point and a lot of other things.  

Adelaide Brown: 

And we have our beloved Kay Siefken leading up, heading up this research, um, projects. So how, she’s talked about how AI is gonna be playing a role. What has she been kind of, um, diving into exploring and terms of chat GPT and GTP and different um, platforms. 

Mike Carr: 

Right, so she’s, you know, she comes from a, a coding background so she knows Python and uh, chat GPT or some of the AI tools out there generate Python code. And so, we have the API. Um, I think any company that’s serious about using AI is gonna get the API and actually do some programming and do some testing with their own data. Uh, where we found it most useful is in analyzing verbatims on research, right? So, when we conduct research, we’ll often try to not get a sense quantitatively as to which names are scoring the best on different metrics. And some of those are system one, some of those are system two, but then the why, right? What is the thinking behind? And, and sometimes, and I think it’s a very honest answer, sometimes the respondent will say they don’t know why, you know, I don’t know why I grabbed this name, it just grabbed me.  

And that boy that is super legit, right? That’s probably the most honest answer anybody can give us. “I don’t know why that name grabbed me. I just like it.” In a lot of cases, they will explain it, well, you know, I associate it with this, or it just is very appealing to me, or it telegraphs these benefits which are important to me. And so when you get hundreds and hundreds and hundreds of those types of, of responses, uh, using an AI tool to go through them and, and sort of recap and summarize what are the themes that are emerging, uh, when we do side-by-side testing, like when we use AI to do that and then we use a person to do that, AI’s getting pretty darn good. I don’t think it’s a replacement for the person, but I do think it’s a great starting point, right?  

That you can, you can run a lot of those verbatims through, you can get some sense as to, okay, here are the themes that are emerging so that when you read the verbatims yourself, which I think you always need to do to understand the gist of the research and, and what the real insights are, you can either valid, you can either verify or say, well, I’m seeing something else that I think chat GPT or the, or the, uh, the Python, uh, model that we are, we’re using missed, right? So, it’s a great tool there. The other thing we’re doing with it is since we’ve been, we have data for naming for 35 years. I don’t know how many thousands of projects we have, um, a variety of data, you know, old creative notes, whatever we may, we may try to feed some of that into the model and see if we can improve its ability to come up with names.  

Based upon a lot of the naming work that we’ve done over the years. Uh, some of that’s proprietary to a client. We can’t use it. A lot of it’s not. And we’ve, we wrote some software tools, in an earlier episode I think I talked about, back in the eighties on both DOS and Windows. So, we’ve, we have roots in writing software for naming. We’ve been doing it off and on for decades. Um, we are fascinated and interested about the new capabilities that AI brings to the table. And so, we’ll probably be playing around with that a little bit more too.  

Adelaide Brown: 

If you’re a new listener, I encourage you to go back to our first episode, first two episodes, and hear Mike talk a little bit more about those AI platforms, some of the first platforms, products, software programs that, that introduced AI, not as we know it today, but certainly as it was in the late eighties, early nineties. Um, well this has been super interesting. I’m excited to dive more into the topics discussed in that “Think Fast, Think Slow” book, um, which we will, in a couple of episodes you’ll hear a little bit more about the system one/system two and how that thought process is rationalized. Um, but thank you so much Mike for, for talking a little bit more about our market research process and um, we’re excited- next week we’ll be talking more about the global naming tradeoff. So, we’ve mentioned that we offer, um, and make sure that our names are screened and available for our clients, but what does that look like when you are expanding from a US domestic product or service to a larger scale international launch? So, I’m excited to dig into that with you.  

Mike Carr: 

Absolutely.  

Adelaide Brown: 

Bye Mike.  

Mike Carr: 

See ya. Bye-Bye.