Tracking podcast data can be difficult. In the past, advertisers have relied heavily on promo codes, vanity URLs, and surveys, and even then, the data would often be skewed leaving advertisers to wonder "Why?"
In episode 20 of On the Mic with Ad Results Media, host Lindsay Boyd sits down with Ad Results Media's very own Andrea Schwarzbach as we chat with Podsights Founder and serial entrepreneur, Sean Creeley. During this conversation we break down the history of Podsights, the importance of podcast attribution, and what the future holds.
Be sure to subscribe to the "On the Mic with Ad Results Media" podcast through iTunes. It's a podcast about the business of podcasting and audio advertising. We aim to educate, enlighten and push the industry forward.
Podcast Transcript
(0s):
Vanity URLs, promo codes, surveys. Now pixel tracking, tracking, and attribution has become practically mandatory in the podcast space in recent years, the need for data and a better understanding of performance has grown as the needs of advertisers who have become more sophisticated. In episode seven of On the Mic with Ad Results Media. We learned a bit about the history of new technology surrounding Podcast attribution. When we sat down with our very own Andrea Schwarzbach and CEO Marshall Williams today, we are excited to bring you a followup episode to that conversation. As we interviewed the founder and CEO of the pixel based Podcast, attribution startup company, Podsights Sean Creeley, Actually buying, and then we move to the reporting side.
(54s):
That's not the problem. The problem is we can figure out what we are growing faster than that. Then I started out with a kazoo. Is that, should I start the podcast off with, could you be so happy? Just the ridiculous to say something.
(1m 35s):
Well, Sean, thank you for agreeing to join us on today's episode. I'm sure a lot of our listeners actually know who you are since we do draw in more of the Audio Advertising crowd. But if you don't mind introducing yourselves for those who don't, that would be awesome. Hi, I'm Sean from Podsights. We do Podcast attribution for brands and agencies. Can you actually break down how Podsights came to be and how it works? Kind of explain it to me as if I have, I've never heard it before. Sure. So Podcasting has a very interesting medium in that it's not radio, but it's also not digital in that we have some digital signals like downloads of user agents, households, that kind of thing.
(2m 16s):
And we have brand pixels that much like Facebook pictures of the Twitter pixels, like what they use on the digital side to do attribution, but it's not sort of the same as the jewel in that we don't have to use their identifiers, but that's what you can do is household level data. So how it came about is that we started the building tools on the research side, right? Like who is actually buying on price. And then we moved into the reporting side, like how many downloads are these agencies getting? And then we, we talked to people and they're like, that's not the problem. The problem is, is attribution. We just, we just can't figure out some stuff works. They, you know, there's a lot of people that have made a lot of good investment decisions and parks marketing very early on through discount codes and the little CPMs and all other fun things.
(3m 0s):
And they figured out if it was working, but as the main sort of the industry expands here, it's no longer a small market, right. It will be a billion and then it will be too. And then they'll be three. And then, you know, sky's the limit kind of thing. And to entice some of these advertisers that we can't use discount codes or, you know, or vanity URLs, or just the tourist inaccurate for a lot of Advertising pixel based attribution's seems to be what works. And that's not the same people don't have success otherwise. But what we have found is that in order to sort of validate that Podcast is working for brands and then to increase scale test with the new networks, with the new shows and the biggest, the biggest attribution methodology works well for its sort of making brands comfortable.
(3m 41s):
I'm with a very different medium. Then some of them I'm sure that this isn't something that you kind of dreamed of doing since you were like 10 years old, you know, starting a Podcast attribution platform. How did you kind of get to the point where you decided to create this? Was there some major event that motivated you? What, what was like your, your background before? Was it an attribution? Was it in the spiritual or tell us a little more about It's in tech startups. This is my second startup. The first one was called deadly. It was in the media space. So we worked with people like Reddit, LinkedIn, and Salesforce and Microsoft to basically build the functionality.
(4m 23s):
Is that like to be paid to Euro into a text box and automatically expands into the video? In general, I like solving problems. I would say sort of building tech solutions to solve them and working on sort of small teams that start something from just sort of the idea phase into sort of a full blown product Podcasting was on our, on my radar. And to be completely honest with you, we were in social recharging. We found that interesting, just watching buddies spend a ton of on Facebook and Twitter and doing absolutely no optimization and just saying, ah, well, let them handle it. That was a bad idea. And we would say like six months, and then I was in a bar bar drinking with a buddy who was like, Hey Podcasting. And he was in the market research and he was like, you know, basically we know that our brand should get into Podcasting, but we have no idea who's even in Podcasting.
(5m 10s):
And then it just starts sort of add to the cascade of following into it, right where you go from one step, okay. Let's just see who's on Advertising and then talk to them and figure out what their problems are and then talk to them, but you can just use and the brands and et cetera. So it took us a solid year and a half to get really any traction. And then once the traction hit and people like Ad Results started saying like, Hey, these guys might know what they're doing. That's when we started to grow. And from a benefit side of things, like what was the decision that Ad Results Media? I mean, why were they like this attribution thing is going to be in the band? Like we knew this is something that's going to be in the industry for a while. Like what, what was the, the change there?
(5m 51s):
Yeah. So I think our agency, as a whole, we've always known that Podcast and radio both probably do a lot better than what we give them credit for or that and what our clients give them credit for. So I think being able to measure that from like an attribution perspective where we can actually have here's the download and here's the person that went to your website and checked out is huge, right? We've never been able to have that kind of data before on Podcasting. It's really this coupon code or a vanity URL. Some of our clients don't even have that. A lot of them just don't have offers. And so they can't do a code, but some of them only measure by survey.
(6m 34s):
And some of them really can't measure on a show by show level, Give you an 80 shows, right. Or a run, a network kind of thing. Like you're not going to have a list of 80 podcasts and say what you want to do here with the software. Well, some, some people do that. We have clients that do that, where they're the dropdown list of shows and you have to go and pick it, Think of the user behavior behind that. Like how dedicated do you have to be doing this show to like scroll, you know, especially with the Z title Podcast. Right. Which is something that we've kind of seen is there are shows that people are such dedicated listeners to that host that they will scroll for days to find that host name, just to make sure that that host gets the credit.
(7m 16s):
But then yeah, there's a lot of other shows where there may not be that listener to host relationship that you can foster through listening to a host over and over again, limited series for example, or it's like shows that a random season, it's a random host. No one really knows the host. You don't get to know them. They're just telling you what the story, you don't really feel any kind of obligation or, you know, loyalty to that host. So you don't really care about going through scrolling through 80 names to find that show. But like, if you listen to Joe Rogan, you want to give Joe Rogan credit. So attribution kind of helps us.
(7m 58s):
And we saw the potential there of being able to show all of our clients that it works better than we think it does. And gives us even more data on a show by show or even a network level of what's actually working and what's driving volume to their site and conversions. Yeah. I mean, it's a somewhat of how we think about as well as this specifically relative like this doesn't make one show, one advertiser doesn't make a whole ton of sense to, to spend Podsights attribution and things with data, right? Because it takes a lot of time, but specifically when you're doing cross publisher or, you know, a large network, it makes a ton of sense to figure out because not all of these shows are created equal.
(8m 40s):
Some are going to work vendors for one, some brands and some just aren't right. And specifically when we start talking about overlap between publishers sort of play and we have done studies here where it's really nice to have them on people saying your name, right. And it's not like Joe, Rogan's sort of the outlier On, we hate talking about him because like, honestly, like he can sell the house yet and people are gonna buy it. Right. But there's sort of the network-wide things where we need to think about like, how do we, how do we sort of scale Advertising here? Because at the end of the game, like our goal was to make Podcast Advertising finger, right? There's still a ton of unsold average check the inventory. There's still a lot of shows in, you know, smaller pumps, Podcast runs that are not getting paid as much as they should be, just because they're basically trying to have to find the advertiser after advertiser and churn, et cetera.
(9m 33s):
So, you know, from a pencil based side of things. Yeah. I like that. It's something we talked about a lot is, you know, optimizing scaling really validating the investment of this space from a more tactical side. I think you've seen this before, but the setup isn't as bad as a lot of people making out to be in the rain. So how sort of attribution works in general is, is, you know, we have integrations with all the hosting providers. So there's what is called the Tracking of Euro for the dynamic assertion. So if we are doing run on network or run the show, it's server to server side, for some of the smaller independent guys, we have a what's called on a to analytics, prefix that host installs sort of a very simple, very easy.
(10m 19s):
And then it's, you know, it's just a pixel on the brand side of it. So, you know, we're not making, which can be done is, you know, smaller shops can be done in a matter of 15 minutes for large shops. You know, whether it's, they can take as a matter of days, but you know, the, the lift isn't that in. And then what we do with basically match households. So it's not attribution it's panel-based right in that we have numerous conversations about how to use versus corporate IMTS versus cell towers and everything else along those lines. But you know, basically 50% of all downloads backpack, a household, and we get, you know, a nice out of there where we can say to someone at this household visit on the brand site after being exposed to him.
(11m 1s):
And then we do a lot of modeling after that to say, well, what does the, the rest of the space, the IP users that we can, that we just can't do any attribution for? How do they, how, how do we think they're going to interact for the site? And that's how we come up with this sort of modeling behavior after that. So it's not some of the survey based ones and some of the discount Euro code ones are, you know, we just apply a random multiplier, ours from a methodology standpoint. Like it's the same, whether it's, you know, one campaign vs the other, which sort of makes it a little bit easier to do some of the validation and what shows are working, but what do we need to sort of rotate out? So you just said, did you say 50%?
(11m 44s):
Yeah. I used to be so like pre COVID. It would of seen a 30, 30%, you know, we can see some, some, sometimes that would be more of the multipliers post COVID. We are seeing a lot of people stay at home and which is good. And therefore sort of the attribution actually gets better. And, you know, things like smart speakers, for example, that are actually tied to the, the household. We have more information then that we can use their, by the way, smart speakers do really well. Right. Because it's, you know, I mean they just buy smart speakers. I'm surprised it's not more than 50% with everyone working from home.
(12m 26s):
Like, is that kind of how it's balanced back out since some places in some cities are going back to work? No, no. We didn't see a come down to like, you know, a hundred percent of people saying as well, right. Because people's sort of French workers, et cetera. There's still a lot of needs going on. There's still people walking around, which is, you know, an outdoor spaces. We, we did a bunch of studies here then around, you know, COVID and downloads like everybody else did sort of, there was a decay and now its kind of pick up and, and all those other fun stuff. But you know, from our standpoint, like we were not trying to shame people for not staying at home and downloading more podcasts. Right. So maybe we shouldn't be, you know, Since launching Podsights you guys launched what a little over a year ago, maybe longer, but I know we started working with you guys, but what over a year ago.
(13m 15s):
Yeah, exactly. It's, you know, basically you've been around since 2018, but you know, in July of 2019, we'd sort of really got going. So what's been the most surprising thing that you've learned since starting Podsights No Podcasting is a mess, right? It's not as simple as, you know, I, I just give Google some money and I give some search ads and, or, you know, or there's a massive display network that, you know, RTB bidding and all those other fun stuff where, you know, I can be on top of the different sites are digital, right? Like it's, it's on a case by case it's a one by one. So most of the Podcasting is people issues, right? It's not actually a lot of technologies and some of the times, so if you're running a campaign across 10 different publishers, it's 10 different could be 10 different hosting companies, which would be 10 different needs with 10 different sales teams.
(14m 8s):
I mean, a lot of it is still like, you know, a RFP gets sent out with people, reply to it. You know, its not as sophisticated as the clients. And there's obviously a lot of pros there, even though the medium is way more authentic. It was a bit so when we went out and, and started building it and we knew, we were basically like, okay, well actually this is all going to be, you know, some big bands for somewhere that is going to go on it and meet, you know, that was probably going to be in and how diverse the space of this it's all over the place. It's not like sort of New York and it may not be subs all over the place, which is super interesting.
(14m 48s):
We are growing faster than I thought we would be honest with attribution in general, is the spaces sort of really taken off in the last year or so. Right. As, as the hosting providers got better as some of the, the, we in a few other people came into the space and stuff like this is actually something that should be solved and the market got big enough, right? You can't have sort of a venture backed business in those smaller markets. And it's really hard if it would have been really hard to basically to start Podsights with VC capital. And because of the fact that like you have to wait a long time and it's like, you know, two years in order for us to get somewhere near where we are and you need a lot of engineering talent in order to make that happen as well.
(15m 32s):
But you know, the, the industry has progressed. So it's very similar in some ways as to how we do digital progress to, and then you had sort of a lot of small publishers and obviously going to be some consolidation you've seen that are on Spotify, Liberty Media, et cetera, but they're still sort of this nice world of sort of the independence and individual publishers that we like to see at least from a buying perspective. Right. It gives you a ton of options. Yeah. And so it really has blown up in the past year. I mean, we have seen other vendors pop up to, so how do you think about staying competitive with everything's still so new with, you know, you guys are still, I think you guys have your dashboard in a really good spot, but you're still constantly changing and evolving.
(16m 18s):
And So we do it right, is that we got to be fairly early. Our whole job was, is not again, the people problems, et cetera. Like a lot of what we invest in as we invest in combination of the partnerships and making sure that when someone comes to the front door, that they know exactly what to expect, that they have a team because it's not like digital where you can just spin up a dashboard. Like there's a lot of things to go through and you know, the, the methodology, et cetera, from a product standpoint. Yeah. I mean, it's constantly just taking feedback and iterating, right. And it's building things for people that you would have thought of them. I think that's what I like the most about sort of building tech startups, right.
(17m 2s):
Is that once you have sort of that initial user base, that the product basically builds itself. Right. As long as you're smart about it, it's taking sort of a, that's the guy who was saying, okay, like how do we not build some one-off for one different client, but make it so bad. We solved the issue, but it works for 10 clients. And then we go sell to more clients and don't figure out what their problems and et cetera. So it's constantly evolving, right? We don't sort of measuring at the same single campaign to measuring multiple campaigns, went to the left to do, and a lot of, you know, overlap analysis and frequency In reach, et cetera. Then nevermind the Ad pulling softened sort of like An allows you, some of the Audio piles back sort of makes the sort of ground up Podcast first.
(17m 46s):
And that's sort of where we have done really well is being Podcast. Some digital providers would come and then give you a big tool and say, Hey, you know, what's going on to that. But you know, Podcast get in front of the fact that like there's no standards and you're integrating with on a case by case basis with a lot of different hosting providers, nevermind like how do we even count to download? Right. Well, I think that's always the same. And mostly because of none of us control the player, which is where, where you have sort of a, a vertically integrated product like Spotify, Spotify, who is the player they own, then the content on there on the ad insertion and et cetera, we can get a little better on the thing, but they just don't have to the scale sort of being on the line thing.
(18m 36s):
Most of the downloads come from third parties and Apple just doesn't care about that. It's the Business. Cause they're a Tracking to truly know what the company and Podcast his surrounding area, but it's a nice app to have on the phone. That's free content. Then anybody can see that mix sort of, of the platform. Again, our job is to sort of work in those, within those constraints and say like, you know, downloads, we don't feel like getting within a percentage point of the impression number is all that important. What we think is the downstream metrics obviously matter. And I think that you guys have done a really nice job of property and lots of clients too, but even back in the discount and surveys, right?
(19m 19s):
It's like gears about the number of persons are getting like how many sales are they driving and how, how much would we be paying for the, for the Media, right. Because of that's positive, we should do more of that. That's negative. We should do less of that, which I think translates really well from the attribution side of the stage. So we try not to get into arguments with people and the side of like, okay, well, you know, it was actually 101 downloads versus, you know, 99. That is not an interesting argument to anybody, right? Like what we care about is that it drove five purchases and therefore, like that was a good investment and we should go and do that again. Just talking about all of the constraints that you do have, what, what would you say has been your biggest challenge so far as, as a company Convincing people that we weren't going to blow up their business, to be honest with you.
(20m 10s):
Right? A lot of people have made a lot of money in this space, like just selling impressions for it and then letting the brand, figuring out whether it was going to work on. Then we had a hell of a time convincing the agencies and the brands and the first place, specifically the agencies, right. Because they're like, well, it's working right now and we are doing well, we don't want, we don't need. And we were still having that problem with a lot of different people saying like, we're not here to help. Right. Are we going to show like that? Every single one of the show converts? Amazing. No, we're not. But we continue to find is that a brand is less likely to churn if they have the information to improve. So even if their first campaign does not do all that, well, you can go back to the whole, you know, dashboard worth of data and say like, okay, well this little part good.
(20m 58s):
Right. And they're used to that, right. They're used to throwing a thousand apps out to Facebook and saying like, this one worked really well. So let's go sort of figure out why it did really well and then double down on it. And I think that's where we'd sort of had our successes, like accidentally falling into one campaign with, with one brand on one agency and then showing them sort of bad sort of cycle. And then watching the brand react to that level of insights vs the spreadsheet after a fact that like this counts. So we got six to six purchases and we are going to give it a multiplier of 3.2 because of the survey, right. That that's sort of a harder sell sometimes to, to make an agency or specifically a brand love the medium we liked dashboard's we were looking at the things as humans.
(21m 44s):
Right. There's just something that we do. So I'm kind of curious just because I'm more on the creative side, over here at Ad Results. Is there anything specific that stands out to you? Like when you're looking at this data, that's like, yeah, that campaign absolutely killed it or like what, what exactly would you be looking for that, that just stands out as like, this was absolutely successful. Here's the problem, right? Like what's where we think the agencies and brands well aligned is that they know the, the, the product, they know the, the, the numbers, they know everything about it. So they can look at what a conversion rate or sort of the CPA, et cetera, and say, this is working or this isn't right.
(22m 30s):
So there's some campaigns that we've done where it's, you know, the Fed's horrible commercials, right? So then the brand like nailed it. And because of the fact that like they're used to getting such low rates on everything else that they do that even though the only drug repurchases, that three purchases just happened to pay for, you know, $3,000 products and therefore, you know, it's, it's ROI positive for good to go right now where we've had the greatest cyber for the most expenses when the brand messaging in the lines of the content, there's obviously no targeting helps for some of the, the dynamic concerted on counts. When you, when you can target based off of either when, or you can't do male female.
(23m 13s):
Right. But you can, you can try to use the things around likes health and wellness, or it decision-makers things like that. Sometimes that works really well. And sometimes it doesn't most of the issues around experimentation and that's where we try to harm ABC creative. Pre-roll, mid-roll, don't buy those roles, you know, make sure the frequency is, is on a reasonable scale. Some people shove you into a bad catalog and you on the frequency of a hundred, I wonder why it's not doing all that well because, because we needed someone downloaded a hundred episodes, but they're only gonna listen. You know, we tell people to buy first friends, whether it be dynamically inserted or embedded, right.
(23m 55s):
The first round is always going to do really well. And generally it's, you can't come in to this industry with thousands of dollars and buy some islands, right. Which sort of is not great for people that are used to Facebook advertising, where you can spend a thousand dollars and see some sort of ROI analysis to be, maybe we need to be in the 10 to 20 to 30 thousands of orders. So you can get a clear sense of, what's not, 'cause what you don't want to do is M 5,000 sort of a back catalog because it's the cheapest thing then, you know, churn because you have no idea whether it worked or not. I asked you the way for you to create a question. I'm not sure if this happens, but you know, obviously endorsements are better than not endorsements, right.
(24m 43s):
Even though we've seen some of the NPR style ads do really well specifically on some of the payroll stuff, because it's the narrator, which is a nice person sometimes where you'll see like the Geico commercials get dropped into the middle of a Podcasting. Like, what the hell are we even doing here? Right. And so, you know, those, obviously we do not work all that well. And then yeah, with increased scale sort of lower conversion rates because of the fact that you, or reaching just a really broad audience versus some of the narrowing that you can do on a show by show basis. So I guess kinda the last question that we have is just what's next.
(25m 27s):
Jerry is going to tell me and some of that's, that's the honest, it's fairly true, right? Some of the times that we're, we're going to see, we continue to see brands that are newer to this space that push us very interesting features that we can add to the product. You know, I don't think that we're going to get to the, you know, it's never like we are going to get more we're on the user. We're just not like we were getting, you're never going to see that those types of insights come in and we're never gonna know who actually downloaded the Podcast. 'cause none of this industry is vertically integrated and that's actually a fairly decent thing. So it's a lot of what we're doing now is, is really just about how do we tell a brand interesting story about how their camping performed and not just show them a insurance because that's, you know, at the end of the day, everybody likes looking at graphs, but then the next question is like, now what?
(26m 28s):
So we have started sort of, we have a data analyst that is dedicated our team. That is the brand can then say like, Hey, needs to give to them with it. We've added some reports for specific that reasons for more touch points to say, this is, this is doing well on this, this isn't doing well on. I think that's important to hear. And then, so, you know, some, the physician side of the side of things, but at the end of the day, yeah. I mean like, what would you want to do is sort of allow people to be more always off versus, you know, do one campaign, spend three months figuring out whether it worked and then trying it again, right? Like you would like to shorten that cycle.
(27m 9s):
So it's, you know, three, three months On, you know, we'd have to look at something and then let's, let's keep on going because you know, this is, this is doing along this isn't meeting. And discovery is obviously a problem that we need to think about. Like how can we help brands even find Podcast of advertise on the Results, obviously as a massive database of all back catalogs of everything that they'd ever done, they can say like, well, for these eight shows, you would expect the same thing. Right. You know, but we started, we started to get benchmarks, right. Which is really helpful for me. And so that when the, you know, the specifically the publisher sales, when I'm trying to sell people into a Podcast, they know like, okay, well this is good.
(27m 54s):
This is bad. Not from a conversion rate standpoint, which will help them sort of make better decisions about what to apply to the hotspot. Mostly Jay is going to tell me No pressure, no pressure. If you enjoyed this episode, be sure to subscribe for updates on future episodes and leave us a comment with your feedback, questions, or ideas for future segments. If you would like more info on Ad Results, Media on what we do, please visit us online at Ad Results. Media dot com. This podcast is an Ad Results, Media production.
