text-talk
* an interview conducted via text message... so excuse all typos *
It's been quite some time since I published a text-talk, and I thought that if I was going to do one, I better make it a good one... and I think this one qualifies!
My guest on this edition of text-talk is Matt Butler, the Co-Founder of Bonsai Data Solutions. Bonsai helps businesses collect and use first-party (1P) intelligence so they can better understand the effectiveness of their growth strategies and marketing activations.
Note: a lightly-edited text transcript of this interview follows the text-talk screenshots below.
A special thanks to Matt Butler for taking the time to text-talk with me and share his insights on the importance of first-party data.
If you want to read more text-talk interviews, you can find them all here.
If you want to ensure you don't miss more content like this when it's published (and you're not already subscribed to dpThoughts), you can subscribe here and have dpThoughts delivered to your inbox every Monday morning.
Text Transcript of Interview:
David Pullara: Matt, thanks for taking the time to text-talk with me today! To start, I'd love for you to take a moment to introduce yourself in your own words.
Matt Butler: Thank you! I'm Matt Butler, I co-founded Bonsai Data Solutions in 2020. We help businesses grow top and bottom lines by designing, unifying and utilizing 1P data. Most clients know us for the intel we bring to their growth strategy -- solving what marketing works and what doesn't. Prior to Bonsai, I was a founding member of Google's analytic consulting arm. I was an "AL" and a tech partnerships leader there until I jumped to Bonsai in 2020.
dp: Well, that gives me a lot of follow-up questions to work with! :)
Let's start with an easy one: Your company is called "Bonsai Data Solutions". When I was a kid, I wanted a Bonsai tree SO BADLY. (This probably had a lot to do with The Karate Kid.) Where did the name come from?
MB: Great question I get this pretty much all the time. The credit for the name goes entirely to my co-founder Mike Remke. We didn't know what to call ourselves and, at first, I wasn't a fan of the idea... But I love it now.
Funny enough, we did NOT name it that because of Karate Kid although everyone thinks we did. The art of bonsai is how we think data strategy can be applied to business. It's about precision, deliberate focus and a dedication to crafting a model of something bigger that's beautiful.
dp: I like that! In your introduction, you mentioned 1P data. Now, I happen to know what that is.... but if I were to ask you to explain the difference between 1P (first-party) data and 3P data (third-party) to one of my young-but-very-tech-savvy children, how would you do that?
MB: Think of 1P data as your story. What did you eat for lunch today? What's your favorite color? These are things you can know about yourself. 3P data are what people SAY about you. Matt likes burgers (I do!). Matt's very serious because he likes numbers (I'm not- I'm a goofball).
3P data isn't always wrong, but it's an incomplete picture. When applied to your business decisions, it can lead folks down a treacherous path. And right now we live in a time of huge amounts of 3P data.
dp: I like your analogy, but I'd like to dig deeper. For example, you can know what I ate for lunch today if I tell you what I ate... but you. can also know if you saw me eating it.
Can 1P turn into 3P, or vice-versa?
MB: Heh great question. So everyone is watching you eat today. And that is a problem because you might not like everyone watching. Also, it's a problem if you make burgers (the business), which is where we try to help.
The 3P data businesses have all told them that Matt came to buy their burger because they watched them. They then ask for more money to go watch more people eat, and then in turn tell the businesses that they deserve more money for watching more people.
When a business is like "but did these people buy burgers BECAUSE you saw them?"...none of the watchers can answer that. And so the business has 3P data everywhere, but no idea which ones to believe.
dp: Does 1P data make it easier to get to the WHY? Like, I bought a burger because I really wanted a burger versus I bought a burger because there wasn't a KFC nearby?
MB: In short - Yes! Here's the beauty of a business's 1P data: when configured/ deployed right, you can know who bought the burgers, what series of events led them to you, and (then) compare that to other people who didn't buy. What events changed? Seeing that difference tells you the WHY. You can also learn about the dynamics - what combination of events has the best outcomes? 1P data unpacks that in a way 3P data never can.
dp: That ties into my next question... our friend John Krebsbach (Bonsai's Managing Partner & CCO, and my former Google colleague) sent me a presentation yesterday to help me prepare for this chat. And while I was going through it, I noticed Bonsai's mission: to "Deliver the future of modern marketing & business intelligence". What does that mean to you, and how does Bonsai help marketers be better marketers?
MB: To me, it's two things: back when digital was still nascent, using 3P data was pretty effective for growth. There weren't many players and 3P data was a lot better than none at all. In 2023, it's completely saturated and hyper-competitive. 3P data isn't meaningful to major marketers anymore. The good news is that any business can operate its own thriving 1P marketing data ecosystem today. The cloud has changed everything. That's the future of modern marketing intelligence. In terms of how it helps marketers be better, that's my second thing...
What people don't talk enough about is that marketing is hard. like, really hard. And often CFOs or others don't value marketers' efforts the way they should because it's been hard to measure using 3P data. A 1P data approach flips the paradigm and gets finance and business to help marketers by getting them the business data that they need.
It's one thing to know Matt bought the burger because of marketing. It's another to know that Matt spends $500 a year on burgers and that the marketers can invest another $300 to win another Matt. This is how it should work. It's how it's always supposed to have worked.
dp: Yes!!!
Hey, Matt, do you know what a real marketer calls "Performance Marketing"?
MARKETING!!!!
(Because all marketing done well is about performance!)
MB: Oh boy, you really touched a nerve there.
I've never understood and always hated that ["performance marketing"] term. It needs to go out to pasture. Performance Marketing implies there's some type of marketing intended to NOT work! It's unbelievable. Marketers need some better branding, don't they?
dp: YES!!!! "Hi, I'd like you to meet Sarah, our Director of Performance Marketing. And this is Bob, our Director of Marketing that doesn't perform." WTF is up with that?!?
MB: I will say it's another issue I'm passionate about, which is to help marketers understand how funny these things are to outsiders. We're all guilty of being in our own bubble sometimes. If the industry could adopt a little more common sense in its lingo, it wouldn't be a bad thing.
dp: Agreed! I think we can also agree that while not everything we do in marketing is easily measured, you can make better decisions when you have data.
You wrote, "The good news is that any business can operate their own thriving 1P marketing data ecosystem today." How? What does a business need to do, in plain English, to collect and effectively use that 1P data?
(And on a related note, what is it, exactly that Bonsai does when it takes on a new client?)
MB: First: you need to understand that "Analytics" isn't something you can buy. It's something you can "do". The main reason a business has bad/no 1P data is that they buy Adobe Analytics or GA or Amplitude and think that they can measure ROI. But customer behavior is unique to your business. These tools can be configured to get the right 1P data, but make no mistake they all need to be configured. That's something Bonsai often times does first. We fix and configure the data going into your analytics so that it is both true and meaningful.
Second: you need to look at all of your marketing spend alongside all of your 1P data and think about 'customers", not tiny actions. It doesn't matter if your cost per click is low, if you look at your cost per new customer is too high! We call that unified customer measurement (UCM).
dp: Can you give me an easy example of something you might fix?
MB: Sure! Your analytics might know your customers filled out a form, but not who filled out the form and what they're possibly worth without that data from your CRM.
We would configure that data in your analytics so you could see that, and know which folks are interested and valuable from the ones that are just interested, but not valuable.
Another example is marketing data ingestion: your analytics doesn't know your Linkedin ad spend from your Linkedin post unless you tell it to read the right details. Having that right might help you grow the right budget/team, and therefore, customers.
dp: Oh, great example! Better data helps you make better spend allocation decisions -- cut back on what's not working well and invest more in what's driving growth!
MB: Bingo.
dp: Bonsai's entire business is built on data... would it be fair to say that A.I. plays a role in helping you build out models for your clients?
MB: I would say so, yes! Al drives our ability to take 1P data and turn it into two things: the "WHY" (we call it Incrementality Modelling), and value prediction: tools we build for marketers that use their 1P to buy the right marketing touch points automatically -- removing their need for 3P cookie data at all.
dp: "that use their 1P to buy the right marketing touch points automatically"... what does that mean, exactly?
MB: We're currently operating what we call "pCV" - a data pipeline that connects seamlessly to Google Ads and allows a business to use Google's Smart Bidding technology (they probably already use it!), but have it run entirely on 1P data. It turns out ALL the clicks you buy have a value, not just the ones 3P data say "converted" (another term that needs to go).
Running Google's bidding Al, but on WAY better data, it turns out that is a pretty good idea. We've seen some great success and believe things like this are really the future.
dp: There's one slide in the presentation John sent that I particularly loved: the one titled "Is your company a fit?" What types of companies aren't a fit with Bonsai and how important is it to figure that out early?
Put another way, what can happen if you don't figure that out right away?
MB: I end up on a lot of strange video conferences when it's not figured out early enough
Sorry, to answer your question...
We see two cohorts really fit: "growth" companies, i.e. VC backed, series A-E, new brands within bigger holdcos, etc. They typically are focused on results from the start and don't have a lot of priors. We are eyes-wide-open in presenting a new approach, but this cohort is fine with that. The second is"enterprise" clients that are clearly in search of a change. They are more complicated, sure, but when there's the right bias to action we are often pretty popular.
We're not great for companies needing to check a "box" if I'm honest. Sometimes this frustrates companies that are used to allocating budget into a specific set of tasks, or companies that don't like their marketing teams and partners to get out of a tactical lane.
dp: Have you ever run into a situation where an enterprise company has a lot of data, but it's all siloed in legacy systems that can't speak to each other?
And if so, is there a solve for that? (Other than throwing everything out and starting fresh, which I imagine isn't realistic or even desirable...)
MB: Every company thinks that these ideas we bring are only possible for companies that aren't siloed and are already advanced.
Here's a little secret: We've ONLY run into that situation. I would surmise the historical trajectory of digital and 3P data proliferation led us here.
All companies are siloed with legacy systems they think can't connect.
That's the trick. It turns out now, you actually can. Sure, there are times to scrap and rebuild, but it's not as needed as you think. These systems can be integrated and 1P Intel can win. We come from the front lines to tell you: it CAN be done!
dp: Well, that feels like a natural wrap-up to this text-talk! Matt, I've taken up a whole lot of your time, so I'll ask just one last question, and we'll end on a fun note...
I know you said you didn't name "Bonsai" after The Karate Kid, but I'm a marketer, a storyteller, and a movie fanatic, so I can't help but ask: if we were to force the analogy, would you be Mr. Miagi, Daniel LaRusso, or Johnny Lawrence?
MB: Oh boy. Mr. Miagi and I have one thing in common. Both vertically-challenged.
dp: You can probably also paraphrase him as it relates to data: "Either you data do "yes" or data do "no." You data do "guess so," (get squished) just like grape." :)
Pat Morita would have certainly said it better, though.
Thanks for taking the time to share your insights, Matt!
MB: Thank you! It was great fun.
<end>