Latest / 1 to 100 / E35: PostScript is more than a tool with Mike Manheimer

Transcript

E35: PostScript is more than a tool with Mike Manheimer · 1 to 100

0:00
Now software companies are being required to drive outcomes and
0:04
that change as opposed to just just enabling A workflow or
0:11
giving people tools and that that changes, creating this new
0:16
issue for SAS companies or something like that.
0:19
And I said it's always been about outcomes and that like bad
0:25
SAS companies weren't about outcomes and good ones always
0:29
have been. And then I think Rashad was
0:32
like, you're come on and talk about it.
0:36
I think that's what happened. That's exactly right.
0:39
Yeah. I just found it.
0:41
OK. Yeah.
0:43
Yeah. What?
0:43
Why do you think that? Why do I think that?
0:46
I mean, I think that like ultimately outcomes are no one
0:52
wakes up at any given day and is like, you know what I need to do
0:56
today? Like I'm today's the day I want
0:57
to buy some software, right? Like that's not, that's not what
1:00
drives people. I wish it did is that's not what
1:02
happens. People like have some sort of
1:06
problem that they're trying to solve that isn't about like just
1:09
clicking the buttons in your software, you know, like for
1:13
Postscript, for example. Like, you know, the, the thing
1:15
that drives our economic engine with our customers is that
1:19
people want to maximize the dollars associated with their
1:24
SMS spend. So they want to like decrease
1:27
their total cost of ownership and Dr. more revenue and more
1:30
sales through through the every SMS dollar they're spending.
1:33
That's why they seek us out or that's what the problem they're
1:36
seeking to solve. Our technology enables that, but
1:40
like if we just deliver the technology and it doesn't result
1:43
in the outcome that they're trying to solve for, then like
1:45
they churn or they go seek it out somewhere else.
1:47
And like I've spent my entire career in software.
1:50
I've literally never done anything else besides SAS in my
1:54
entire life. It's the only jobs I've ever had
1:56
were SAS companies. And like, I feel like that's
2:00
always been the case. Like if you are a with a winning
2:02
SAS company, it's because you're driving an outcome for your
2:07
buyer that like solves their problem.
2:10
As precious as we all probably want to be about how cool our
2:13
tech is doesn't really matter. The only reason people renew is
2:18
because they are like, I came here to make more dollars or
2:22
gain efficiency or whatever it is that that that the value prop
2:26
is that they were seeking out. And like, if you don't deliver
2:28
that, then you're getting cut. And I've never worked in a SAS
2:32
company where that wasn't the case.
2:34
And I'm talking about SAS. One of my first job in SAS ever
2:38
was in 2009 when like SAS was like barely a thing.
2:45
Like literally we're competing with on Prem solutions.
2:47
So it's like day one of software.
2:51
But it's always been that way, you know, it's always been
2:53
outcomes is how you're measured. And if you want to be really,
2:57
really successful, you have to deliver those.
2:58
And so my contention was like, this isn't new.
3:01
It may be AI is making the the way that you have to do some
3:05
things different, but like ultimately it's always been you
3:09
deliver outcomes or you end up creating churn for your
3:11
business. In my.
3:13
I couldn't agree more. I used to be a customer of
3:15
Postscript back when I was an agency and I remember when we
3:17
were looking for the SMS product that we were going to use across
3:20
the agency with, you know, all of the different brands or over
3:23
100 brands that we onboarded, we had a conversation with
3:26
Attentive. And because you had to have a
3:28
conversation with Attentive to start using Attentive and the
3:31
conversation felt like, well, what are you going to do for us?
3:34
Or the sales guy was like, you know, really just trying to
3:36
maximize the sale on it. I'm here to solve a problem
3:39
about SMS. Like we ended up going with Postscript because it
3:42
was like log in, get in, get out and the job's done.
3:46
Yeah, I mean in some ways like you know a lot of us are are
3:50
marketers or marketing adjacent, right.
3:52
Like there's like, you know, the jobs to be done framework and
3:55
you know, we can don't need to talk about whether that's the
3:58
right one or the wrong one. But like ultimately it's like
4:00
kind of pointing at this, which is like there's a, there's
4:03
something that a customer's trying to hire you to do.
4:07
And if you're very aligned to that and you deliver that thing,
4:12
then like that, you know, we, we like to make everything really
4:14
complicated, right? It's pretty simple in some ways.
4:18
And if you can be very aligned to that job, you understand your
4:21
customer well enough that you like, feel like you know what
4:23
that is and point your whole entire company at that, then I
4:28
think you, you create not only strong outcomes for customers,
4:31
but strong outcomes for your business In, in my opinion, that
4:34
underpins all of the successful SAS stories that there are,
4:38
whether you're like bootstrapped or venture backed or whatever.
4:41
I, I don't really think it matters.
4:42
I've been in a bunch of different environments.
4:45
I've always said like I think the person who's closest to the
4:47
customer is going to win. And when I say closest to the
4:50
customer, I mean I understand the customer problem as deep as
4:54
the customer, so I can act accordingly.
4:56
And acting accordingly to me has always been, I'm going to solve
4:59
that core problem in a way that's like defensible and that
5:02
the customer agrees with. And if you do that, then you get
5:05
multi year renewals, which we all know we want and love.
5:09
If you don't, they'll go find something else.
5:12
The difference though, is like, I think the expectation of what
5:17
software should do has changed a lot.
5:18
So it used to be that software just had to enable you to do
5:22
something. Now software is being asked to
5:25
do the thing for you. And I think that's a, that's a
5:29
gradual shift. But yeah, Mike, I think the way
5:33
I would push back on even Postscript is you have an agency
5:38
running Postscript for a large number of brands and the brand
5:44
gets their outcome by hiring an agency to operate Postscript for
5:49
them. And the reason they do that is
5:52
because of things like thinking about the campaign, thinking
5:56
about the content, thinking like someone has to do the work to
6:01
get the outcome. Yeah.
6:03
But like that's only, I would actually push back on that a
6:06
little bit. We are just just to
6:09
contextualize what she was saying.
6:11
We have a branch of Postscript called Postscript Plus and it is
6:18
essentially what he's talking about.
6:19
There's it's it's more or less an agency.
6:21
It's like a hand over the keys. We do it for you sort of
6:24
service. Some of our customers use it,
6:28
some don't. The the amount of customers that
6:30
use that is like sub sub 15%. But some customers need need
6:41
that to get the outcome that they're seeking.
6:44
Some do not because they have really great in house retention
6:48
teams who know what to do and are going to like partner with
6:52
us to deploy a strategy that's going to like help them achieve
6:56
their goal. Now our sales team, if in the
6:59
sales motion they get a sense that the thing that the client
7:03
is seeking out is unlikely to be achieved with just our
7:07
technology and their in house team, they recommended this
7:10
other service and we find a way to make the economics work.
7:13
But ultimately we only prescribe that if we think it's critical
7:18
to the outcome that the person's trying to get.
7:21
And so and like, that's just a professional service which has
7:24
existed like in every software business model there's ever
7:27
been, you know, SAS on Prem software before, there's always
7:31
been professional services to help help people get what they
7:35
want out of it. No.
7:36
No, no, dude. I this, see this is where I
7:38
think there's a nuance that matters, which is that I think
7:43
we are moving closer to a world where customers will expect
7:50
their software vendors to solve the whole problem.
7:57
Just like in the previous world between the professional service
8:02
and the software or the team and the software or the agency and
8:06
the software, That's how you solve the whole problem.
8:10
I, I think you will end up in a, now we can call them whatever we
8:16
want, right? We can call it for deployed
8:18
engineers. We can call it professional
8:21
services. We can at some point automate a
8:25
large majority of it with agent workflows like it doesn't like
8:31
the the verbiage I think is less relevant to what is the customer
8:36
thinking. The job of the software vendor
8:38
is and I think that the total degree to which the customer is
8:44
expecting the software vendor to solve the whole problem is
8:48
shifting like that. That I that I think is is likely
8:52
right. I'm not, I don't contest that
8:55
actually. I mean, but I think that that
8:57
what you're, to me, what you're describing is just people are
8:59
asking for an expansion of the services that we offer.
9:03
You know, like they're the the the the problem is expanding
9:07
that they're asking us to go attack.
9:09
I agree with that. And I mean you can see some
9:12
really good examples of that in the market where people are
9:14
adding on adjacent services like, you know, the financial
9:16
services companies are good examples like the ramps of the
9:18
world, the ripplings of the world that are becoming like
9:21
they're doing everything for everyone, but they're being
9:24
that's just an expansion of their their model, what their
9:29
customers are asking them to do. And people ask us to do more.
9:34
People are probably asking Vermont to do more.
9:36
I'm I'm not contesting that that's the case.
9:38
I'm just saying that like the idea that SAS one point O
9:42
companies got away not delivering outcomes for
9:45
customers is just like in my opinion, not true.
9:49
I think what Rishab is saying is that the expectation isn't
9:53
actually that tied to the idea that they want an outcome is
9:57
that they don't. They want an outcome without
9:59
having to engage a service as well or service arm.
10:03
It's like, and, and whether that exists in the product right now
10:06
or not. I think it is like one of these
10:07
like gradations of, of expectations that is moving into
10:11
a, into a space where you're getting exposed to like tools,
10:15
even if it's a different, different category that
10:17
seemingly end to end do the thing that you were trying to
10:20
achieve with like one or multiple people powered by
10:23
software before. And I think that the expectation
10:26
of value, especially for things that potentially cost a little
10:29
bit more, you know, like if you can draw a straight line between
10:33
how much your thing is costing and then salaries.
10:35
I think what is happening is that people are expecting the
10:38
software to replace people. And again, I think it is like a
10:43
graduation depending on on what their experience was with this
10:47
like category of software service before.
10:50
But I just think it's that the expectation isn't particularly
10:53
on just like the outcome. It's that they want the outcome
10:55
by paying for the software and not engaging with the service,
11:00
or at least less of a service. Right.
11:03
I think it's an interesting overlap between what Taylor was
11:05
saying last week, though. People do value people right now
11:10
more than they value software. Paying $2000 for software seems
11:14
expensive. Paying $2000 for a person is a
11:16
bargain, right? So, you know, that service
11:19
component of Postscript seems smart.
11:21
That is something that people value really highly, and it can
11:25
be competitive with labor. Maybe it's a way to, you know,
11:29
earn pretty large deal sizes and solve the problem more
11:33
comprehensively than what is possible with software right
11:36
now. Yeah, I guess like what we're
11:38
talking about too is like the what the what's real now versus
11:42
what's the future? Like are people asking for this
11:46
right now? Like I, I think that the easy
11:48
example is I think probably what's happening and I, I don't
11:51
know this first hand. I don't work in for a a company
11:54
in customer service, but it seems like the fins of the
11:57
world, the real value prop that they're being asked to solve is
12:00
like I want to decrease, I want to extract costs out of my
12:03
support org and I will give dollars to whoever can can do
12:07
that. Whether or not that is going.
12:13
Obviously they've been growing that business really
12:15
effectively, whether or not that is good for the customer.
12:19
And I think it's probably dependent on the deployment and
12:21
in any business, but we'll set that aside.
12:23
I would say like in a lot of places, and I'll just speak for
12:26
what I see everyday. Like people really aren't asking
12:30
us that yet. Like, people are asking, people
12:36
are coming to Postscript being like, I'd like to extract human
12:42
capital dollars out of my business to run my owned
12:47
marketing channels. And to do that I would need to
12:50
hand you the keys. In fact, actually we get when
12:55
that conversation even comes up, it's mostly the opposite where
12:58
people are pushing really, really hard back on that,
13:02
especially if it's meant to be done by AI Now.
13:07
I think like obviously we're early in the life cycle and
13:09
eventually we think that maybe the capabilities will get to a
13:13
point where where brands might be like more interested in that.
13:17
But right now, I think it's like a glimmer in people's eye, and
13:23
they're not asking for that today.
13:26
This is this. Is this OK?
13:30
I'm going to get in trouble for saying this but I'm going to sit
13:32
this is this is because of how organizations are built and has
13:36
nothing to do with people's actual preferences.
13:39
The reason why the support team is the first one that is being
13:45
articulated in this way is how many brands do you know where
13:49
the support team is sitting next to them on desks next?
13:52
Like my desk is here and then the 20 support people desks are
13:56
next to me. 0 OK, they're all outsourced.
14:00
They're out of sight, out of mind.
14:02
There's no ability to organizationally push back.
14:06
And that is why that is like the single best place to make the
14:11
change. First, it is an organizational
14:14
behavior problem. It's not A and desire problem.
14:19
It's and desire. It's also the easiest spot to
14:25
get results and see those results immediately.
14:27
Because of this. It's much simpler than a lot of
14:29
these other spaces. And I like when you have
14:33
documentation that AI can reference to, to understand
14:36
things. You've got all of this context,
14:39
you can feed it in a way that you can't typically without the
14:41
balls. And I think that's a very
14:42
important. Part of it as well.
14:45
Yeah, Jeremiah, I, I agree with that.
14:47
Like, there's a lot of other consequences of an
14:51
organizational setup where you hire third parties that need to
14:56
respond off of a script, right? The script has to exist, for
15:00
example, right? And so I'm not saying that there
15:05
aren't other benefits of this organizational setup, but the
15:09
organizational setup is effectively the underlying way
15:13
that this thing, right? Like that, like these things
15:16
feed off of each other. And so then you don't have the
15:19
organizational pushback when someone shows up and says like,
15:22
Hey, don't you want to cut your support team by 70% like that?
15:28
I mean to, to, to say like, hey, it's not going to happen in
15:32
other departments. Is, is, is, is only an artifact
15:39
of organizational setup. It is, it is not an artifact of
15:42
the UN desire. Like there's not a business
15:44
owner out there who does not want to be able to get to build
15:49
their business to be bigger with less people.
15:52
It's less communication overhead.
15:54
It's less cost like like that, that that is like for sure what
15:59
people want. Right.
16:00
If you think it's separately separate from the efficiency
16:02
angle and like unlocking scale, you can imagine a world where
16:06
you know, more automation and particularly SMS marketing would
16:10
actually be more effective, right?
16:12
Right now, if you rely on people, maybe they have certain
16:16
biases, they have a small sample size of experience that they're
16:20
using to educate their decision making, but you have access to
16:22
all of the data and you can get data in theory from the Shopify
16:26
store. See, I don't know what is most
16:29
likely to drive a second purchase.
16:31
Recommend the right products based on their first order,
16:34
recommend the right products that maximize lifetime value.
16:36
All of these things that on a, you know, send percent basis,
16:42
especially if you include all the segmentation that would need
16:44
to happen would just be massively out of scope for a
16:46
human right. So totally from efficiency and
16:51
measurable. You're you're, you're preaching
16:54
what we're what? You're taking the words right
16:55
out of our mouth. I agree with you 100%.
16:57
I actually think like our road map is very tied to this idea
17:05
that we believe we can create an environment where we are
17:11
significantly more likely and we can prove it to you via a
17:14
holdout test that we can find in the right next best message to
17:18
maximize your SMS dollars. Ultimately, that's the the value
17:22
prop people seek us out for. What I'm saying is that like
17:28
it's maybe in in support because people view that as a a cost
17:33
center, right? There's not expectation are at
17:37
least less or so on performance there.
17:39
And like Rishabh said, they're not sitting there.
17:42
So they are in a position where they probably can't defend
17:44
themselves the same way. I think that what's in our way
17:50
or or in any performance channel that's measured that way is way
17:54
to being into a full automation situation is yes, organizational
17:59
pressures are a component of it, right?
18:02
There's people there that you're working alongside.
18:03
The structure impedes the decision to hand the keys over
18:08
to full automation for sure. But also the performance is I
18:14
think the the priority, like when you can demonstrate at
18:18
scale that you can outperform it in in against whatever the
18:24
status quo is will facilitate that change.
18:28
And like, I don't think a lot of people believe that today, not
18:32
not talking about post group search knowledge in general that
18:35
that's possible. And of getting back to where we
18:39
started, I would say it's because it, it doesn't achieve
18:42
the outcome that they want, which is like to maximize the
18:45
SMS dollars. Right now people I think would
18:47
think, well, the guy who I've been working with for 2 1/2
18:50
years, who's rebuilt my own marketing channels from scratch,
18:53
I still believe that person can do a better job running this
18:57
machine than you guys could. We one day believe, we'll
19:01
probably, like Jack was saying, be able to do it better.
19:04
But until they believe that they're putting their dollars
19:07
and effort to where they think they're going to get the best
19:09
outcome and we have to act accordingly.
19:12
And like, yeah, we're going to have to be threading a needle in
19:14
the, in the adoption curve for sure.
19:16
But I still think it all like comes back to like, what are
19:20
they trying to achieve with the software?
19:21
How do they believe they're best going to achieve that?
19:24
Are we showing up for them in a way that helps them do that in
19:28
the best way? And whether that's a fully
19:29
automation AI world, whether that's with professional
19:32
services or whether it would be like fully PLG self-serve, you
19:36
know, it, it, it depends on what outcome they're seeking.
19:40
And like that the being able to react appropriately to that
19:43
outcome is like, I think if you're a SAS person, like that's
19:47
your job. The puck's moving.
19:49
You have to be able to move where the puck is going.
19:51
But like the job's, the job, which is like, what does your
19:53
customer want to achieve? Can you deliver that in a way
19:57
that is commensurate with the value you're asking them to pay
19:59
for your software? And it creates a a business that
20:02
like can exist for yourself. And you know, sometimes the puck
20:06
moves in a way where those things become incompatible.
20:09
Sometimes the puck moves in a way where it actually
20:11
strengthens your ability to have a have a strong SAS business on
20:15
top of that. But ultimately, like we're all
20:17
trying to like meet the outcome needs of the customers.
20:20
And sometimes that requires people, sometimes it doesn't.
20:22
I think in the future we think it'll require less people.
20:26
But the only way to know that is to deliver the outcome, to prove
20:30
it right. And so, like, it's always about
20:33
the proof. Yeah, there, yeah.
20:36
The other thing though, is what you're talking about, it has the
20:38
potential to massively expand the use cases and the way in
20:43
which your product is used and the complexity of how it could
20:46
be used, right? So there's a world in which that
20:48
doesn't mean less people, it means people being more
20:51
effective with the tools they're using.
20:53
And I think that's something that that's something I think
20:55
about a lot with the way that we're thinking about AI and our
20:57
products is I think the easy pitch is when we replace people
21:00
for, you could say it's a line item you can look at and say,
21:02
OK, yeah, this makes sense to pay $2000 a month for this thing
21:06
versus $20,000 a month for those people.
21:08
But the more powerful thing in the long term is actually how do
21:11
you expand the ability for this to drive better outcomes for
21:14
your business? And that, I think, looks like
21:17
enabling people to do more impactful things.
21:21
Yeah, totally. I mean, we think, you know,
21:24
it's, it's there's for a similar proxy to like the way that like
21:27
Meta has worked overtime where it's like the, you know, media
21:31
buying to create a strategy, right?
21:33
Like I don't want our customers trying to like think of the
21:39
best, most complicated segmentation strategy.
21:41
Like we're, we're either there now or very close to being able
21:46
to do that algorithmically way smarter than than they could.
21:50
Not because they're not smart, but because it's like a human
21:54
can only, yeah, a human can only have visibility to a certain
21:57
amount of variables in a really complex system.
21:59
Obviously a computer can do that better.
22:01
Now the like actual hook, the thing that goes into the text
22:07
that drives the engagement or that would show up in an ad, for
22:10
example, That stuff, I think it's probably pretty far away
22:18
still of us being able to do that better than, you know, as a
22:22
traditional like retention marketer.
22:25
And so like the jobs are going to change for sure.
22:27
I'm not sure. I think there's probably some
22:29
overlap here with, you know, the paid media stuff that you're
22:31
talking about. I think the the reason why we've
22:33
seen up until more recently and and maybe still now so much
22:38
terrible ads, so many, so much performance created, it doesn't
22:43
live up to the expectations of the brand and the brand looks at
22:47
it and goes well. If it works, it works is because
22:50
the medium is just so good, right?
22:52
Facebook is so good at showing your content to the right
22:55
person. It doesn't matter if you have
22:56
some shitty ad, it still might print right?
22:59
And I think that maybe, you know, I see this in clay and
23:02
Postscript and retention platforms as well.
23:05
You have your performance dashboard and it's doing
23:07
hundreds of thousands of millions of dollars, even if you
23:10
just turned on the template. Welcome flow, right?
23:13
Why? Because you're sending messages
23:15
to people who are very likely to purchase, right?
23:17
So I wonder if the medium is just so effective the that a
23:21
baseline, especially for those people who are looking for
23:24
efficiency as opposed to I want the best outcomes possible.
23:27
They're looking for. I want this thing to operate and
23:29
not to pull my hair out and have to talk to one more retention
23:33
person every week about, you know, what changes they're
23:36
making, just actually being able to do all of that.
23:40
If that were the criteria, Postscript would be
23:43
significantly bigger than it is because you're totally right.
23:47
Like there's like the I, I'm curious what you guys see in
23:51
your customer basis. But like that, I think that's
23:53
not, that's why this is tough because it's not solely does
23:58
this perform better, right? Like I can send the like most
24:03
click baity text on the entire planet.
24:05
That's like the thing that you looked at 10 seconds ago is
24:09
about to be out of stock for the rest of your life in about 8
24:12
seconds. And I'll give you an 80%
24:14
discount if you click this button right now.
24:16
And like, I could like be like this performs.
24:19
But as you guys know, I think because our customer bases are a
24:22
lot of overlap, 99% of brands are like, just because that
24:26
outperforms what I was going to do, I'm not willing to do that.
24:30
And so the calibration of like the mix of brand, which is like
24:34
everyone believes that, you know, brand might be like the
24:37
last real mode. And whether you, we can debate
24:40
whether you think that that's real or fake and that's just an
24:42
illusion to, to people. But like, you know, people are
24:46
in this because they're trying to make a brand, they're trying
24:48
to balance performance, which is actually super critical.
24:51
And how those things come together and show up in, you
24:54
know, a, a medium that is part creative, part performance is
25:00
like, why this is difficult? And the amount of times I will
25:03
show people something, I'm like, I guarantee if you do this,
25:07
it'll generate you more ROI. And then they say, well, how do
25:10
I ensure it's on brand? And then we get into that
25:12
conversation about like, well, what does that mean to you?
25:15
And where are the guard rails of that?
25:16
And how does your team, you know, like and all that stuff,
25:20
you know, that's the that's art and science and that that part,
25:24
at least today, is still art, you know, so and especially in
25:28
the way that you interface with clients on it.
25:30
I'll give you the most generous view and then I'll give you the
25:32
most provocative view that that I have of this debate.
25:37
The most generous view is what is I think what you said,
25:40
because you're a generous man. And and so like, I think that it
25:45
is, it is possible like the one end of the spectrum is that,
25:50
hey, there's a set of things that the human intuition is very
25:56
good at that we have not yet been able to coach a machine on,
26:00
OK. And we can people use different
26:03
words for this. Some people call it taste, some
26:06
people call it brand, some like, and then there are just people
26:09
who are exceptional at that. OK.
26:11
And, and for sure, I agree that there are people who are
26:14
exceptional at that. Do I agree that every marketer
26:19
who operates every one of these platforms is exceptional at
26:22
that? I think everything lies on a
26:24
bell curve, right? That's the most likely outcome.
26:27
The most likely outcome is that everything lies on a bell curve.
26:29
OK, so that, so let's just say that's one end of the spectrum.
26:36
The other end of the spectrum, which I think is a more
26:38
provocative 1 is what if the machine could do it and could do
26:44
it better. And the the problem is the human
26:49
psyche, which is like, we hate the loss of control, like the
26:54
like we want control and we want impact, which are highly related
26:59
to each other. And so, Yep, even if I think
27:03
that this like, even if I have all the evidence in the world
27:06
that this thing is going to do better, if I'm not the one who
27:09
thought of that, like the whole, you know, God, what's the
27:13
Leonardo DiCaprio movie where you like, embed an idea?
27:20
What an assumption. Inception.
27:22
Exactly. Unless I can get all the way to
27:24
Inception, I'm like not going to convince you to like pick the
27:27
thing right And and like. Those are the 2 ends of the
27:32
spectrum. Well, it depends on what time
27:34
horizon. Like I, I think like eventually
27:40
it'll get there, right? Like eventually it will like I
27:44
think, well, maybe we should ask this like, do you believe that
27:48
eventually when I say like, you know, 20 years out that like
27:52
it'll give it the longest view possible?
27:55
Do you think that the machine will be able to be coached to do
27:58
that thing correctly? Yeah.
28:01
Well, I guess what I'm saying is I think there are places where
28:03
it is already better. And because it's not reliably
28:07
better, we still have to fall back on the human because of
28:11
this lack of reliability. And so I'd actually think that
28:14
the the time frame within which this will happen is extremely
28:17
fast. Like I, I, I think like roughly,
28:21
you know, I'm like a simple man. Like when Zach or Sam Altman say
28:24
something, I just sort of like, listen to what they're saying
28:26
and I'm like, OK, that's probably true.
28:28
And so Zach said that by the end of next year, he's going to be
28:31
able to like automate meta. And so I'm like, that's probably
28:37
true. You know, like he has said, a
28:40
bunch of stuff that has roughly been correct.
28:44
Do you remember that metaverse thing and there was like jungle
28:49
noises and everyone was going to be wearing goggles and living in
28:52
the metaverse? I do remember that.
28:55
I mean like it may, I'm not trying to say I think it's one
28:58
way or the other like whatever, but I think what you're talking
29:02
about is like you're changing behavior, right?
29:04
And how long does it take to do that?
29:05
Like regardless of what the performance is, like there's
29:08
some timeline that we are going to have to get through, even if
29:12
the technology is there, that's going to take time.
29:17
But that's, that's just like, you know, how I I don't know, I
29:20
mean, at any time you're like doing something new, whether
29:25
it's driven by a technological leap or whether it's just simply
29:28
a novel approach to an existing category of software, for
29:32
example, you know, there's some, there's some adoption curve.
29:36
And I think, you know, maybe short term it'll take longer
29:39
than we all want. But I think it seems like we all
29:43
kind of have a sense of where we're going.
29:45
And the question is just like how fast will happen?
29:47
Nobody knows. I wouldn't bet against Zuck
29:49
either, Rashaab. So I get what you're saying.
29:53
But yeah, I think, I think in the near term, you know, is is a
29:58
question mark. In the long term, it seems like
30:00
an inevitability. I think direction and taste will
30:04
be like the only thing that's left because like if let's just
30:08
say meta ads are fully automated from end to end and I'm selling
30:14
like this microphone and there's like 5 other companies selling
30:17
the same microphone with the same specs.
30:22
Like the only thing left then is like direction from my company
30:26
and brand. And if at the end of the day,
30:29
it's just like a new status quo that that gets created and like
30:34
most companies can't survive just like at the status quo, or
30:38
at least their ambitions don't align with, just like we can
30:41
sell a certain amount of microphones to a certain subset
30:44
of people, Like the mandate is going to be, we need to sell
30:47
more microphones than we did before.
30:50
And that means selling more microphones than like our
30:52
competitors. And so therefore, we now need to
30:54
have a stance or a taste or an angle in which was not created
30:59
by the machine. It just reiterates the thing
31:04
that was always true, which is like, brand matters a lot.
31:07
And over any time horizon, any sort of technological or product
31:13
feature gets copied and and multiplied by a bunch of other
31:17
people. And that's just going to happen
31:18
faster and faster and faster and faster until the end of time.
31:22
OK. Like what direction is this
31:24
company actually going to go in? Dude, totally all I'm suggesting
31:30
is that the total number of people who can do that is not
31:32
large. What why is it that certain
31:36
brands are so like our create the culture or create the psych
31:43
like and those are the ones that endure, right.
31:46
So it's like Apple is thing different.
31:47
Nike, everyone's an athlete essentially, right?
31:50
And they actually set culture. And but the total number of
31:55
people who can do that is not large.
32:00
And so for the rest of us, you know, like including me and, you
32:05
know, it's like I'm just going to hand it to the machine.
32:09
Them, it's way harder to say. Every single person has a
32:12
different point of view and they want us as software providers to
32:15
be able to deliver something that will consistently reflect
32:18
their point of view regardless, regardless if it's the best or
32:23
the highest performing or if it's like the cultural
32:26
tastemakers. That's really tough because
32:29
that's all up here. And I also agree with the Zach
32:31
who's saying like the people who are saying that's what they want
32:35
us to do, they believe in their head that there's like some
32:38
defensible thing about that, right?
32:41
Like they they agree this is the last defensible thing.
32:44
This is my Moat. So I need you to hit that.
32:48
That means that you know that if your ICP is the top 100,000
32:53
stores in Shopify, you're being asked to develop a creative
32:56
model that can reflect back to the merchant in 100,000 distinct
33:01
ways reliably, instantly, their brand, not the number one brand.
33:06
That's way, way, way harder to do consistently I think, which
33:10
is why people are like, I'm not handing the keys totally over
33:13
just yet. Well, also like Nike as an
33:18
example, didn't define culture by defining the masses at the
33:22
start. And no culturally defining brand
33:25
actually ever did that. Like, and then Apple's like
33:29
another really good example. It's just like it was a bunch of
33:31
fucking weirdos in a garage and they would meet up and hack away
33:35
at stuff, you know, like they, and it's like, like another
33:39
example would be like a technology company would be like
33:41
a company. I really like, like Teenage
33:42
Engineering that make like really random like audio
33:45
hardware and like things like that.
33:47
It's like they, if they took the playbook from Apple, they
33:52
wouldn't be successful in the mission of potentially defining
33:55
culture in like 30 years. And so I, I actually think that
34:00
if you did like lean too much into like AI stuff, you know,
34:04
you could end up losing your shot if your ambitions were to
34:08
become like a culturally defining brand.
34:12
Because like, part of it is like being super like stringent about
34:16
niche. And you can't really have a
34:19
large enough opinion when you're not big enough.
34:21
So you kind of have to pick like a smaller thing to have an
34:23
opinion about. Yeah, I could agree more.
34:27
I think if we think about culture representing demands,
34:29
then it can be responsive or reactionary to what the market's
34:34
asking for. As a SAS company, that might be
34:36
understanding the way that people are engaging with
34:39
software that's extremely popular right now and what
34:41
they're asking from it. You know everything about me
34:44
make the best recommendations or take the best actions possible
34:48
on my behalf. But I can agree more with you,
34:50
Zach, that I think culture and supply side, it might be brands
34:53
big and small with vision defining what it is that they
34:57
think how things should be better, right?
35:00
So I could be a small brand, right?
35:02
Saying that for my customers, for people like me, this is the
35:06
solution that I I think should exist, right?
35:10
It's actually interesting when I really think about this, this
35:12
actually exists currently in the way the brands are distributed.
35:16
There's lots of little micro communities that brands have
35:21
that brands serve. And in those contexts, you don't
35:24
see competition really, because the reality is the market's not
35:27
big enough to have much competition and it takes a lot
35:30
to execute against that. And so I don't know, like came
35:35
to think of a good example off the top of my head.
35:37
Here's the one I saw the other day, A company that makes, what
35:42
is it like a like fermented kelp products?
35:46
Like it's just like such a niche thing that like maybe that could
35:51
become a big category someday, but there's basically 1 brand in
35:55
that category, right? And the, and I don't even think
35:58
they're doing that well. But point is like there's a
36:00
bunch of middle categories like that all over the place.
36:03
And then you've got like T-shirts and you can see a brand
36:05
like true Class that come in and in three years become, I mean,
36:11
there's still not even that big in that category.
36:12
But like that's a massive success to be able to go from
36:16
zero to hundreds of millions in revenue in a few years.
36:18
And you've got a Grun's supplements, massive category,
36:21
daily, daily vitamin, basically massive, massive category.
36:24
You can make a bunch of noise very quickly.
36:26
So I almost wonder if it looks like people with a unique
36:29
perspective have the ability to move faster against incumbents
36:34
and incumbents have the ability to protect themselves
36:37
potentially against those moves a little bit faster.
36:39
But ultimately it kind of looks like it does today where you got
36:43
less competition in small categories, more competition in
36:45
bigger categories, and everybody's trying to go after
36:47
whatever that pie is that that makes sense in that moment.
36:51
Yeah, but I could tell you that if you're trying to provide, you
36:54
know what I'm talking about, like from an SMS lens, you know,
36:57
we all do different stuff. So like whatever the creative
37:00
output, right is component of it, right.
37:01
You, you have a software, there's the performance side
37:04
there, there's the workflows and all that sort of stuff.
37:06
And then there's the brand side that has to you put the creative
37:09
into the thing and get the distribution.
37:11
I can tell you with certainty, though, this was why this is is
37:15
really challenging that like if you created like a creative
37:21
prompt that was would share the between true classic and cuts
37:28
and built and viori and the the the and the fresh clean Tees
37:32
that the T-shirt industrial complex.
37:35
If you were like I, I've nailed that right.
37:38
They would all be equally unhappy with the output.
37:41
It has to be theirs has to be reflective of them, even if it
37:46
performed really well. And so like, I think that and
37:49
then you can go like a level factor down.
37:51
Like I was just thinking about like I was talking to the
37:53
founder of born primitive recently and like technically
37:58
they sell like leisure wear shoes, shirts, like they're
38:02
they're clothing brand, right? But like that's even a level A
38:05
factor level down. Like it's very specific.
38:07
It's like veteran owned fitness oriented.
38:12
They're targeting a very specific type of person.
38:16
But like technically they sell, you could say that they're part
38:19
of the T-shirts that get sold in Shopify too.
38:23
It's just, it's not like for like, right.
38:26
And historically, I think in SAS, like you would say, I'll
38:28
print this once and sell it to a bunch of people.
38:31
And as you know, folks who are building an ecom where creative
38:35
is in brand is a very, very big component of it.
38:40
You can't really just do that. Like the bar is way, way, way
38:42
higher. And I think it's, you know,
38:45
getting back to the outcomes part of it, like that's still
38:48
part of the calculus that the brand's using on whether or not
38:51
they're going to buy and renew our, our software or, or hand
38:54
the keys over to our AI or not is if you're, if you're hitting
38:59
that. And it's, it's a really tough
39:00
thing to hit because the true classic and built and cuts and
39:08
Viori and fresh clean Tees are not trying to like hire the, the
39:16
generically best T-shirt salesman over text messaging
39:20
that they can get. They're trying to hire the best
39:25
person that's selling T-shirts over text messaging for their
39:27
brand specifically and only that.
39:30
And you know, that's, I think a criteria of, of that's, I think
39:40
it, it's everywhere, but very concentrated in ECOM just given
39:43
the nature of what we do. These are all entrepreneurs who
39:45
have a really strong point of view about why their brand
39:47
should even exist in the world in the 1st place.
39:49
And they're honoring that constantly.
39:51
It's different for every single person, which makes it tough,
39:54
which is why I'm saying that like today, do I think
39:56
eventually people will be asking for this?
39:58
I think they will. I think today, unlike what
40:02
happens with support agents, which are like cut these costs
40:04
out of my business, people are like.
40:07
I don't believe you that you can nail that for me specifically
40:11
better than the team that I've hired and cultivated over the
40:14
last many years. And they're probably wrong on
40:17
the segmentation side and they're probably currently right
40:20
on all of the taste brand, etcetera, the stuff that they
40:24
consider to be their MO. I have a few questions that I
40:27
want to like to to push on this idea a bunch and and take them
40:33
in whatever order. One question is what is the
40:36
economics of that asked differently?
40:41
Like what is the actual value as expressed by willingness to pay
40:46
of getting that outcome? Second related question is
40:51
articulate for me why someone does or does not sell on Amazon
40:57
where you are AT shirt seller and AT shirt seller is AT shirt
41:01
seller. That's like the most extreme
41:04
example of the sort of take away all uniqueness and like, hey,
41:11
you're your T-shirts and then articulate for me, how does that
41:16
brand think about being inside of physical retail?
41:22
So like, let's just pretend for argument's sake that in what you
41:27
are describing in the digital world in your own channels as
41:31
possible, then I want to understand the economics of
41:34
that. Let's then go to the next level,
41:36
which is Amazon, which is a little bit like closer to you
41:39
are just another person in this category.
41:42
And Amazon tries very hard to make you look like another
41:44
person in this category. I know that's not exactly true,
41:48
but let's just say that. And then let's go all the way to
41:51
retail, which is like, you're in this aisle, right?
41:54
And I know, look, don't get me wrong, I understand.
41:56
That's why packaging is such a big topic, because everybody's
42:00
thinking about when you look at the shelf, how do you stand out?
42:03
But the retailer is not thinking about that.
42:05
So, so, so, so just let's deconvolute those things as a
42:09
way to interact with this question a little bit more
42:13
tangibly. On the economics side, the first
42:16
part of it, are you saying economics to us or to the brand
42:19
or both? Sorry, the the what is the full
42:23
economics of it? Like, I don't care who who gets
42:26
what economics like I want to understand what like how much
42:30
value is getting created. Like economics to us is just
42:33
what value capture is happening, which is like not a particularly
42:36
relevant thing. It's like, help me understand
42:39
the value generation that's happening for the economy.
42:43
And then of course, how that gets split is like a business
42:45
model question. So like not not actually that
42:47
interesting, right? Well, I, I don't know.
42:50
I actually think that there's something that could happen as a
42:53
result of, of nailing this that's incredibly unique to SAS
42:57
companies of the future. That's which I'll talk about in
43:01
a second. But like, so you're saying
43:04
what's the economics of being able to like nail the
43:06
performance side and the brand side for a company to be able to
43:12
to sell the most shirts in their own way?
43:15
Yeah, presumably total value has to go up, right, Otherwise you
43:19
wouldn't do it like the economic system doesn't work if that's
43:23
like the in the idealized outcome, total value has to
43:27
increase, right, other otherwise like the system doesn't work,
43:30
yeah. Yeah.
43:31
I mean, well, I mean, one way to look at it would be like in all
43:34
of us in some way work on this, right, Which is, you know, the
43:38
the standard conversion rate drop off is like the thing that
43:42
we're all trying to help people solve in different ways, right.
43:44
Like you buy a certain amount of traffic, people go to the site,
43:48
a bunch of them don't convert, some of them do.
43:50
If we can get more of them to convert, the pie gets bigger.
43:53
We all influence that in some way theoretically.
43:56
And I'll just use, you know, the SMS example because that's what
43:58
I'm familiar with is theoretically if we could create
44:02
like the best T-shirt salesman for true classic over SMS is we
44:08
would increase the conversion rate of people who who buy after
44:14
being subscribed to the channel. So there's some implied number
44:17
that that is theoretically we get so good at that that we
44:20
could drive that number up to it's like natural, like perfect
44:23
equilibrium, which would be somewhere higher than it
44:26
currently isn't true. Classic would be making more
44:28
money, which would, you know, not only be making more money,
44:31
but would be like decreasing the cost of ownership of the channel
44:33
because theoretically we wouldn't have to scale costs at
44:35
the exact same amount. So that's how they would make
44:38
more money if we like nailed that.
44:40
The thing that I think is incredibly interesting for SAS,
44:45
right, because the the thing that we're probably all dealing
44:49
with in inside of our own businesses and every every SAS
44:51
company deals with this like churn is the big thing, right?
44:54
If you have a too leaky of a bucket, you can have a
44:56
successful SAS business. The business model that came
44:59
before us, which was on Prem had against the customers will a
45:04
very sticky element to it. The servers were under the desk
45:07
right, You own, you know, they releasing those like you had to
45:10
show up there to deliver the thing that's pretty sticky.
45:15
I think in this new world with these new tools and AI and what
45:18
not, we might be headed into an environment where we're getting
45:22
the stickiness back right because of true classic beliefs
45:27
that Postscript has cultivated the best possible salesman over
45:32
SMS for them. Then like theoretically that's
45:35
better lock in than anything that we've had in in SAS before.
45:39
It's getting as close to the the old way, which was, you know,
45:43
very sticky. So I actually think like there,
45:45
there's a reason, obviously there's always a reason to do it
45:48
to help customers make even more money.
45:50
But that level of specialization, I think creates
45:54
new, less portable features of your software and your service
46:02
delivery to your customers that might make them stickier than
46:05
they would naturally be just doing a good job today.
46:08
Because if you build taste in these things that are, you know,
46:11
we're talking about how, you know, amorphous these things are
46:14
in the 1st place. If you nail that on behalf of a
46:16
customer inside of your software, like theoretically
46:20
that'll make you significantly more valuable, be harder for
46:23
other people to copy. And so I think there's economics
46:28
on both sides that are really interesting and make the whole
46:30
thing worth pursuing. But ultimately, no one does any
46:33
of this. If I can't make true classics
46:36
conversion rate go up and more revenues coming in because we're
46:39
increasing conversions in a way that doesn't scale at the same
46:43
rate as their costs. If I can't do that, then I don't
46:46
have anything. I have to be able to do that.
46:47
That that's really the thing that they're buying from us and
46:51
we'll be rewarded appropriately. I feel like there's potentially.
46:55
A bit of a misalignment that when we think about like the
46:57
macro versus the individual brands and that like, I think
47:00
for the consumer, the ideal outcome of AI and like the
47:04
personalization ability is that it creates less waste.
47:09
Like in my family, there are things like like let's say I buy
47:12
a true classic T-shirt. I don't like it, right?
47:18
I, I buy this T-shirt, I'm not going to return it.
47:21
It goes in my closet. I wear it a couple of times.
47:23
I try another brand, I try another brand like you kind of
47:26
and even within brands, obviously this can happen too,
47:28
right? But the ideal outcome isn't
47:30
necessarily like macro wise isn't necessarily that true.
47:33
Classic sells more unless it's actually a superior product to
47:38
the other products that are out there, right, Which I I think is
47:41
probably the case. That's why they're growing so
47:42
fast, right? They it's clearly superior to a
47:45
lot of brands, but I think the challenge here is that it's not
47:48
like I think the, the actual best outcome maybe is less stuff
47:52
being purchased from some places, more from others.
47:57
And you're actually like at the like macro and the level and the
48:01
individual consumer level. It's actually just like a more
48:03
efficient process. I just think that's in conflict
48:06
with individual brands sometimes.
48:08
So not to say that would be with true Classic, but there's some
48:11
brand out there that's losing in this context or has to lose in
48:13
order for that to be. You're you're saying because the
48:17
consumer leverages AI to like make a more correct choice and
48:20
the and that means that the value tends to accrue to the
48:24
best product in the space. It should, yeah.
48:26
I mean, and at the end of the day, like no matter if we can
48:29
drive more sales for brands that you can't necessarily drive more
48:35
spending power in the economy at that level, right?
48:38
Like there's a, there's a cap to that.
48:40
There always will be like it, it goes up over time, but at a
48:42
certain level, there's not an ability to exceed whatever the,
48:47
the aggregate spending power is. So I, I, I guess I, I, I don't
48:51
know how that plays out. Like the incentive will be for
48:54
AI to be inaccurate sometimes in terms of what's beneficial for a
48:58
consumer from the brand standpoint at least because you
49:01
want to make more sales. It's not necessarily going to
49:03
be. That would only be that would
49:07
only be true if there was literally no differentiation.
49:11
They get that there was a complete commodity right?
49:13
Like. Yeah.
49:15
I guess what I'm saying is 2 brands competing at each other,
49:18
let's say they're 5050 market share, right?
49:20
They both want more sales. They're both like both, both
49:25
brands are incentivized to sell more.
49:27
But the actual for the consumer, that's not actually what's
49:30
beneficial. And so I don't know how that
49:32
that plays out. I, I, to me, I don't know the
49:34
answer to this. I'm just like, it's an
49:35
interesting thing to think about.
49:37
Like in an optimized world where AI is driving a lot, what does
49:40
that actually look like? Does it look like more
49:42
purchasing? Does it look like less?
49:44
Does it look like math Winters and lifters?
49:47
And at least a price compression because, Mike, have you ever
49:49
tried to ask ChatGPT how it makes product recommendations?
49:55
You, anybody listening to this at the 59th minute should do
49:59
this. They should go to ChatGPT and
50:01
say, can you tell me how you would like what criteria you're
50:05
using when you make a recommendation for a product #1
50:09
price. And then you ask it, why did you
50:12
pick price? And then it says, that's a good
50:15
point. I don't know if you care about
50:17
price and and it's like this very funny interaction actually,
50:22
but yeah, you're going to have all of this conflict all over
50:25
the place, so. You do assume over time though,
50:28
that with personalization to you, whether it's predictive or
50:32
it's because of your past conversations, that you start to
50:35
get recommendations that are like sought from high to low,
50:39
right? And ideally, you start to get
50:41
those recommendations without even having to have a
50:42
conversation, right? You don't have to say, hey,
50:45
actually, I want the best, not a compromise, right?
50:49
It just knows that about you because it's integrated enough
50:52
into the rest of your tours. Yeah, I don't know.
50:56
I think it's interesting. Yeah, I mean this their context,
50:59
right. Personal context, like all of us
51:03
are have T-shirts, right. And I'm sure if we all went into
51:06
each other's closets, we all have different brands, right.
51:08
There's like context of why we all made those decisions.
51:10
It's personal to every single person.
51:13
Yeah, there's. So I think that they're in a is
51:17
the context is important because most like it's not surprising to
51:23
me that you would say like, OK, choose some criteria to sort
51:26
these price being #1 because that's like a piece of
51:30
structured data and it's a sortable on every site.
51:33
So like that must be important. So here's price.
51:36
But theoretically there would be a world where it would know you
51:39
or be able to take data from other systems to contextualize
51:42
the selection, right? Like the, because it's very
51:47
clear in our economy that not that's not the thing that
51:50
everyone cares about most, you know, the the Isai Miyake Apple
51:56
thing is sold out, right? Like, you know, so like price is
52:00
not the driver of 100% of the dollars in the economy.
52:04
Yeah. Especially for retention.
52:06
It's, it's not, it's not at all like the, the reason why I would
52:10
buy AT shirt again is if I got complimented on it or
52:15
potentially my, whatever my, my experience when washing it was
52:19
bad or something. But like, and that's where,
52:23
like, that's what I'm saying, like brand, brand matters.
52:26
Like if someone puts an ad in front of me, that's like, and
52:32
the hook is, I've never been complimented so much in a
52:35
T-shirt or something. Like I doesn't, I don't give a
52:38
fuck what Chachi BT has to say about the blend of cotton and
52:41
polyester and I don't think anybody makes purchasing
52:45
decisions like that. And like I know for myself, like
52:48
I, I, I have like brands that I just like for no reason.
52:52
Like one of the craziest ones is like James purse.
52:54
It's like the stupidest price for a T-shirt.
52:57
But I buy them because I get somehow makes me feel something
53:01
that like nobody else knows and like chat GP would never know
53:04
that. And so like I, I, I actually,
53:09
I'm not a huge, I, I think for buying software like Chad GBT is
53:12
like really good, but I don't think for like buying something
53:15
like AT shirt like Chad GBT is going to make that much of an
53:19
impact is specifically on the retention piece of buying the
53:23
2nd, 3rd and 4th T-shirt. If anything, it's going to
53:26
introduce reasons why I should try something new.
53:29
We find a lot of times I make suggestions on what you else you
53:33
should buy from the portfolio of products, which is really
53:37
valuable to to the brands that are you know what hero skew, but
53:41
have a huge long tail of things to like deep in the relationship
53:45
and the connection by providing subscriber context in the mix
53:49
too. Or they might be like, you know,
53:53
imagine it knows Zach, you moved to from a, you know, warm
53:58
weather place to a cold weather place or vice versa.
54:01
It might make different types of suggestions that help drive you
54:04
deeper into the, you know, the brand ecosystem.
54:09
One day, hopefully automatically today, you know, you'd have to
54:15
like have a good retention marketer who like, thinks of
54:19
angles and like tries to build the segments to go target the
54:21
people who might be doing those types of things.
54:24
And so I think eventually that's where it's going.
54:30
How fast we'll get there is is a question mark.
54:33
Well, this has been good. Thanks, Mike.
54:36
Yeah, Thank you, Mike. This is an interesting
54:37
conversation. I got to tell you, I want to see
54:40
it. I want to see that, you know,
54:41
check a box and and do nothing in Postscript.
54:44
But we might have to wait a little.
54:47
While we're we're working on it, there's, there's, there's the
54:52
development of the technology component and then there is the
54:56
adoption of the technology component.
54:59
Both equally important. They are related, but they're
55:03
not the same thing. Yeah, yeah.
55:05
Hey, dude. Just like with Shopify itself,
55:08
the adoption is obviously going to come first from the people
55:10
who are launching new brands because then there is no change
55:14
management, right. So that's that I think is
55:17
actually the, I mean we didn't get to talk about it, but that's
55:20
actually the most interesting thing is I think it's completely
55:22
understated that actually all of this change in software, the
55:26
single biggest thing it's going to do is create the next wave of
55:30
new brands. Every technology cycle has
55:33
basically done that. And yeah, that I think is
55:36
actually going to be the most interesting and it's going to be
55:38
the people who are native to that way of working that end up
55:42
being tomorrow's jewelry and true classic and whatever else,
55:47
right? Just like these guys were native
55:49
to Shopify and Facebook ads. Yeah.
55:51
I think the, the one thing that we didn't touch on as well is
55:56
that to some extent, the new brands that are going to fail
55:59
because they pursue efficiency, they use AI before it's there
56:03
yet and they they get lazy as opposed to making the best
56:05
product. They're going to exist, right
56:09
where they could have one if they went with taste, right?
56:12
And that might be the same context of Postscript, right?
56:15
Those brands that I don't know, using ChatGPT and just copy
56:18
paste straight over and bundling the bag, right?
56:22
They're going to exist right in the service of just automating
56:26
everything. Right.
56:27
And I've been, I've been actually surprised a little bit
56:29
at how many people push back against that immediately.
56:33
Like I thought more people would just be like, oh, cool, feed the
56:36
machine and we'll just see what comes out of it.
56:40
But in my experience, at least in this, you know, year one year
56:44
1 1/2 with like commercially available AI products available
56:49
to folks, the quality, taste, brand guardrails, whatever you
56:56
want to call those are they're there already.
57:01
And people are actually not they're they're excited about
57:05
the potential, but they're they're tolerance for what
57:07
they're willing to accept. Like what's like the AI slop
57:10
tolerance and retention marketing is actually incredibly
57:16
low. And it's it's much lower than I
57:20
thought it was going to be on, you know, try number one.
57:23
I thought, you know, maybe people would say, you know, 8020
57:25
is totally good. What we're finding is like, no,
57:28
it's like 99 one. And you know that it's our job
57:34
as a software vendor to like figure out how to hurdle that
57:36
now. Yeah, going getting the 90s
57:39
easy, getting to from 90 to 99, it's really, really hard.
57:44
Awesome. Mike.
57:45
Thanks for coming on. Yeah.
57:46
Thank you. Anytime.
57:49
Bye guys. Catch.
57:50
You a minute.