Anthony Kennada sits down with Tim Sanders, VP of Research Insights at G2, to talk about why good data is the only sustainable competitive advantage in today's competitive landscape.
0:00
Well, Tim, first of all, congratulations on your new role at G2, VP of Research
0:04
Insights.
0:05
You know, G2 is sitting on such a treasure trove of data.
0:10
Last I pulled up here, 2.5 million verified reviews, 90 million annual visitors
0:15
who are
0:15
researching products on G2.com, and over 160,000 products and services are
0:21
listed.
0:21
I'm curious as you're thinking about approaching this role, can you just speak
0:26
to the power
0:27
of data and insights in original research as a function of a content marketing
0:32
program?
0:32
Well, to quote one of my favorite artists, Wu-Tang Clan,
0:36
data rules everything around me, right? So when you think about what's powering
0:42
artificial
0:42
intelligence, it is data. And this is especially true with the rise of gener
0:47
ative AI, right?
0:48
Because generative AI allows natural language, like what we're doing, to become
0:54
the interface,
0:56
to all of the world's data. However, the outputs are only as good as the data
1:02
is high quality
1:03
and irrelevant to the situation. The other thing I think about when I ponder
1:09
this concept of like,
1:10
how important is data? Because we say it's existential. It's the most important
1:14
thing,
1:14
but I'm going to put it in a different kind of light. So when you think about,
1:18
you know, technology today where AI is becoming always included more than artificial
1:25
intelligence,
1:26
when you think about it, right? What's happened is that for the first time in
1:30
human history,
1:31
technology, software, has decoupled prediction from judgment.
1:36
First time ever, in the past, prediction and judgment happened at the same time
1:44
in a human being called an expert. And they set up, based on my experience,
1:49
I think this is going to happen. So I prescribe this action. And oftentimes we
1:52
would integrate it
1:54
into a firm with a rule of thumb. And the problem with rule of thumb is that
1:57
they have
1:58
invisible expiration dates. You really don't know that the rule of thumb doesn
2:02
't work until you've
2:03
had your third quarter in a row of things falling off a cliff. And then you're
2:06
in a doom loop. So
2:07
what happens is if we let the machines make predictions based on high quality
2:12
data,
2:14
then the humans simply pass judgment. And there is an opportunity now to scale
2:19
personalized decisions for the enterprise, right? Every situation now you're
2:24
going after it with
2:25
code or codified solutions. So it creates an incredibly dynamic organization. I
2:30
think about
2:30
ant financial in China, it's a Harvard business case, fastest growing financial
2:35
services company
2:36
in the world. They got to 500 million users with literally 10% of the employee
2:42
base of Wells Fargo.
2:44
Wow. Because they had decoupled this thing. The analogy I would use here is
2:48
that when you think
2:49
about the data behind ways or Google Maps, it's profound. The quality and the
2:54
quantity of that data.
2:56
Why is that important? Well, I'll give you an example from a recent trip.
2:59
Which the United Kingdom, Love London, always loved my London cab experience.
3:04
There's always
3:07
been a shortage of London cabs. It created huge queue times during rush hour
3:11
and surge. And the
3:12
reason why is the industry thought they had a logistics problem. I've got to
3:16
buy more cabs and
3:17
I've got to recruit more drivers. No, the problem was they had a prediction
3:20
problem.
3:21
You had to go to school for almost four years. It's called the knowledge.
3:26
Before you could be a
3:27
cab driver for London cab because you needed to learn every possible route
3:31
scenario so you never
3:32
had to pull over and look at a map. Because writers hate that. And as a result,
3:37
it created
3:38
a huge crimp in the funnel. So we're like we've seen now with GoToMarket. And
3:42
then along comes
3:43
ways. And along comes Uber who thinks differently and says, "Really, all you
3:47
need is human judgment.
3:48
So let's put an app in their hands." And if they know not to honk and not to
3:51
tailgate,
3:52
not to have music on, not to take calls judgment, then they can be a great
3:56
driver. And now you've
3:57
got a thousand percent more drivers in the United Kingdom. Mobility has been
4:01
restored with
4:02
artificial intelligence. It's just invisible. So my message is data is your
4:08
only mode. For any
4:10
organization, there is no other sustainable competitive advantage. Well, let's
4:15
stay on that
4:16
topic of AI. Well, we're here. You referenced Harvard. You're recently
4:20
appointed executive fellow at
4:21
the Digital Data Design Institute at Harvard. And my understanding is this is a
4:26
function that
4:27
enables researchers to deliver insights that drive the adoption of AI more
4:32
broadly. And I'm
4:33
curious of just, you know, we're here at this Better Together conference
4:36
thinking about the
4:37
application of really we just referenced the mobility kind of exercise that
4:42
what ways is introduced.
4:44
How is AI going to shape these emerging use cases for the GoToMarket functions
4:49
within businesses as well?
4:50
So to pivot from D-cubed, that's what they call it at Harvard, D-cubes research
4:55
has been very
4:56
use case focused. They've also looked at governance. They've looked at AI
4:59
safety. I'm looking a lot
5:02
at packaging use case work they've done to kind of meet their charter, which is
5:07
to democratize
5:08
digital transformation and AI for all of business people. Right. So I think it
5:12
's a fantastic purpose.
5:14
So I get behind it. I'm the flavor of evangelist called explainer. So I'll talk
5:18
about that in a
5:19
minute. So I want to pivot to McKinsey research because this is really
5:23
interesting.
5:24
McKinsey did an analysis at the end of last year and it was on point to gen AI,
5:29
which is the fast
5:29
motion right now in AI. And what they looked at is they looked at what
5:34
functions in an organization
5:36
can you have the highest revenue impact first and second were sales and
5:40
marketing.
5:40
Right. Okay. Now, it doesn't have the highest take out of cost center, highest
5:45
take out of
5:45
cost center is customer operations. Doesn't have quite the revenue impact and
5:49
it's got a lot more
5:50
cost behind it up front. But in the long term, it could take out cost. But go
5:53
back to sales and
5:54
marketing. Incredibly high revenue opportunity, low hanging fruit for an
5:59
organization. Why is that?
6:01
Because when you think about prediction and then you think about automation
6:05
based on prediction,
6:06
letting the machine now take an action on your behalf or following the machine
6:10
's advice.
6:11
It's only as good as your tolerance for hallucinations. Sure. Right. So the
6:15
last thing we'll ever see
6:16
is autonomous driving. The first thing we saw was cadence improvement with
6:21
products,
6:21
right AI, writing assistants like Grammarly, right? Because we have a tolerance
6:25
for it being slightly
6:26
wrong because we can fix it right with the human in the loop. So sales and
6:29
marketing of all organizations
6:32
more than pricing, more than R&D, certainly more than customer operations can
6:36
experiment and they
6:37
can fast fail and they can figure out what that production model looks like to
6:40
take advantage of
6:41
prediction and not collapse the business. So that's why immediately they are
6:47
the functions. Go to
6:48
market where generative AI can produce incredible results and perhaps build a
6:52
moat behind it.
6:54
Yeah. It's super interesting. I have to ask the off script a little bit, but we
6:58
've leveraged,
6:58
we've used this language in B2B or Go to Market for some time of marketing
7:02
automation,
7:03
which almost feels like in the 20 year context in which that term came up,
7:07
mistimed or too early for its time to some extent. But now, as we think about
7:14
predicting
7:15
actions using data, potentially, agentic workflows to take the action on half.
7:21
Covering that. I'm excited about the agentic economy. Yeah.
7:24
Very curious about like, is marketing automation, true marketing automation now
7:27
coming?
7:28
Well, humans, be in the loop, have to be in the loop given some of the lack of
7:34
tolerance for
7:36
hallucination, maybe for at least a high profile prospect or whatever. Just
7:40
curious, where do you
7:41
see the marketing mix headed? So first of all, when you think about the
7:45
disciplines of artificial
7:46
intelligence, there's two. There's machine learning, which is rather mature. It
7:50
is very simple.
7:51
Zeros and ones. That's the way I talk about it. Very simple. Yes, no, this,
7:55
that kind of predictions.
7:57
And then there's natural language processing. The artist known as Gen AI, much
8:01
more complicated,
8:02
much more sexy, because any fool can get value from it. M.L. is a black box. As
8:06
a result, machine
8:08
learning has become the nickelback of artificial intelligence. Even though you
8:11
know it works,
8:12
you would never tell your friends at a party. It's a private pleasure for data
8:16
scientists.
8:17
Automation right now should really be called marketing velocity, because that's
8:24
what you're
8:25
really talking about. You're talking about increasing the velocity of go-to-
8:29
market motions that capture
8:31
the value of innovations. That's what we're really talking about. Automation is
8:36
a path to
8:37
velocity. The issue though is that there's automation, semi-automation, path to
8:43
automation,
8:44
I think right now with Gen Ritavaya, we're still at path to automation because
8:48
it has a high human
8:49
and the loop requirement. Automation is only a goal in as much as velocity has
8:57
impact on your business.
8:58
You have to ask yourself, faster to market is that an advantage for us. It
9:03
could be.
9:04
The other thing is getting back to this McKinsey research. For sales and
9:08
marketing, what they
9:09
looked really hard at was the idea that when you're selling, and this is
9:12
important for the sellers,
9:14
when you're selling artificial intelligence, you've got two models to sell it
9:19
to the customer.
9:20
You've got the replacement model. They call that the cost model. I'm going to
9:23
bring down your cost
9:24
measured by headcount reduction. Then you have the capacity model where you say
9:30
I'm going to augment your existing team members so they can increase production
9:34
without increasing
9:35
headcount. Very sexy. What they say is organizations that sell based on the
9:40
capacity value proposition
9:42
will create more satisfaction, find more true value, and our terms get better
9:46
renewal and expansion.
9:48
Why? Because we're never satisfied with the substitute. That's the inherent
9:53
flaw in looking
9:54
monolithically at a concept like automation. The idea is that AI is not going
10:00
to take your job
10:00
and kill your company, but humans that are armed with AI that strategically
10:04
focus will
10:05
that take your job and kill your company? Totally. Well, you know, I want to
10:09
just coming back to
10:09
your role at G2. A big part of it, as I understand, is evangelism as well. So
10:14
obviously doing the
10:16
research, but then getting the word out to market, maybe doing conversations
10:19
like this and several
10:20
others. Evangelism as a concept has been around for a while, but it feels like
10:25
over the last few
10:26
years we've seen new terminology around things like creators or influencers
10:31
really start to
10:33
pop up as well. I'm curious, how do you think about evangelism as a function of
10:39
helping brands
10:40
build these relationships with a market? And maybe how does it differentiate
10:44
from some of these other
10:44
emerging functions as well? A creator produces content. An influence who
10:50
produces results.
10:51
The influence, that's the difference. There's an old Chinese proverb,
10:58
if they're not following you, you're just taking a walk, right? You're not a
11:01
leader. So an influencer
11:03
inherently grows their following in a measurable sort of way. They grow the
11:08
experience of people
11:10
taking their advice and putting it into action. So I think a lot about
11:14
connecting with influencers.
11:15
One of the things I'm doing for DQB at Harvard is we're syndicating insights to
11:20
certified digital
11:22
influencers to spread the word beyond what we can do with say Harvard Business
11:26
Review or whatever.
11:28
Let's talk about evangelists. I got interested in this concept maybe 25 years
11:32
ago when I was
11:33
working from Art Cube and Guy Kawasaki, by all accounts, would be your first
11:38
codified chief of
11:39
evangelist for Apple, right? So he was a mobilizer. What he did is he mobilized
11:47
the emotion of design
11:49
excellence and he transferred it from Steve Jobs Mind to the end users where
11:54
they said,
11:55
you know what? Computer's personal computers should be beautiful. They should
12:01
be able to be up and
12:02
running in five minutes and they should be simple and they should almost
12:06
disappear into the fabric
12:08
of our lives instead of being this thing that a Hewlett Packard or Gateway
12:12
sending us. So his
12:13
job is to mobilize people to make decisions based on design and not price. And
12:18
as a result, when you
12:20
think about the effect of Guy Kawasaki and we're here in 2024 to quote Prof G.
12:24
Scott Galloway
12:25
my favorite podcast out there markets, Apple has become the number one luxury
12:31
retailer in the world
12:32
who now sells scarcity, right? We never think about the price of an iPhone. It
12:37
is ridiculously high
12:39
because we think about the design opportunity, right? So when I fast forward
12:44
because that was
12:44
exciting to me, I fast forward just a few years when Yahoo buys broadcast.com
12:49
and I go to Silicon
12:50
Valley. I work in the value lab, which is kind of what I'm doing at G2. We're
12:54
converting data into
12:56
actionable insights to increase leverage for the sales team and get clients and
13:01
keep them.
13:02
I really had my eye on being the next Guy Kawasaki and I got very lucky because
13:07
our CEO at the time
13:08
felt like we had a what you would call a go-to-market problem. No one believed
13:13
in digital advertising.
13:14
The trope was banner ads didn't work and all of a sudden market is say we're
13:19
going to pay only
13:19
on clicks. Well, we ship impressions and that should be rewarded but there was
13:23
a real problem in the
13:24
market associating us with the kind of impact of television printer radio. So
13:31
now we had to sell
13:32
on clicks which could take profit out of the model 80, 90 percent. So the job
13:36
of ValueLab
13:37
was to create data but what we lacked was something public facing to make it
13:42
simple for people to
13:43
understand. So I became Chief Solutions Officer which was a proxy for Chief
13:47
Evangelist and my job
13:49
is to get in front of all of our significant quality opportunities and serve as
13:53
what I call an
13:54
explainer. Okay, so there's two kinds of evangelists right now. And I'm just
13:59
making this up. There's
14:00
what I can tell you about. There's the explainer that takes the mystery out of
14:04
something and
14:05
democratizes it for the listener. The listener feels empowered and the listener
14:10
usually becomes
14:10
an evangelist. Okay, then there's the mobilizer. Mobilizer connect and this is
14:15
really based on the
14:16
corporate executive board research from Challenger customer. The fantastic
14:20
follow-up to challenge
14:20
the sale. The mobilizer knows how to get people within their organization to
14:26
embrace change and
14:27
they have the gravitas to unseat incumbents. Okay, mobile hours are powerful.
14:32
They don't just
14:32
have influence. They have power, right? So that's a different kind of evangel
14:37
ist because that
14:38
evangelist, their job isn't to democratize the product and the technology and
14:43
the innovation
14:44
because the problem for them is in mystery. Their job is to do a change
14:47
management because there's
14:49
some kind of scaffolding at the customer level that's inhibiting utilization
14:53
which shows up with
14:54
where's the value. Right. You know, because I believe right now for software
14:59
the biggest problem
15:00
when it comes to renewal or perceived value is that we don't as sellers take
15:06
ownership for
15:07
utilization. Right. We put it on them. Yeah. Right. We do a couple of cursory
15:12
moves. So right now the
15:13
mobilizer's job is to convince everyone in the organization through the
15:18
acquisition of mobilizers
15:19
and they become evangelizers too. Change is critical. There's a burning
15:23
platform. Change is easier than
15:26
you think with rapid collaboration and change is going to be your competitive
15:31
advantage because
15:32
once you learn how to be agile and adaptive learning effects, think of it like
15:36
network effects on
15:37
steroids, learning effects will create a win for you that no one can catch up
15:42
with. So those are the
15:44
two types and they approach different problem spaces based on the enterprise.
15:47
You don't have to be
15:48
both. Yeah. But I found in the last year is that if you're selling artificial
15:53
intelligence,
15:54
especially you're not selling Nickelback, you're selling this new kind of AI as
15:58
a higher cost
15:58
structure, probably a higher on ramp that you have to suffer through and you
16:02
have to have a higher
16:03
tolerance for mistakes because it's not as clean as ML. Right. If you're
16:07
selling that, I think you
16:08
need to be an explainer. Interesting. Okay. Right. So I find that when it comes
16:12
to getting people to
16:13
the table with their checkbook and CFO right next to them, just democratizing
16:20
that AI is just a
16:21
prediction machine. Yeah. And that's all it does. It takes information that you
16:26
currently have and
16:28
produces data that you don't have. Yeah. Just that bit can unlock C suite who
16:35
currently see it as a
16:36
threat, see it as a bad right, see it as something that is currently overpriced
16:41
. We're going to wait
16:42
for the bubble to break all of those really come from that internal ego pushing
16:47
back and says,
16:48
I don't get it and I'm a smart person and that bothers the crap out of me.
16:51
Right. So
16:53
that's what I think the most important motion is for AI. However, what I think
16:57
a little bit
16:58
downstream and I look at at software solutions that you define as being in
17:05
competition with the
17:06
status quo in that situation, I think you need a mobilizer. Gotcha. Right. So I
17:11
think you need to
17:12
accordingly authors of a challenger, customer, Matt Dixon in particular, you
17:17
need a dog whistle
17:18
that brings the mobilizers out and that's where content marketing, getting
17:22
connected with revenue
17:23
organization and product and success where all of those functions coming
17:27
together can
17:28
absolutely bring mobilizers out so we can create relationships with them. We
17:33
can enable them through
17:34
a variety of different tools to go back home and sell, especially to all the
17:38
decision makers we
17:40
can't round up and talk to. So I think you have to make a decision what your
17:43
problem is.
17:44
Are you selling something that's just new and confusing or are you selling
17:49
something that
17:50
requires some internal change and behavior for them to utilize and realize
17:57
value and then you should
17:59
go pick your evangelizers wisely. Right. That's so good. Well, look, we're one
18:04
time for one last
18:05
question here. We're here at the Better Together event and I'm curious.
18:09
Obviously, you're ramping
18:10
up in the AG2 in this new role. We're even at Harvard or as you kind of think
18:14
about the work
18:15
required to educate, to build the insights and so on. What are two kind of
18:21
tools that you leverage
18:23
to technologies, to services that work better together in your own work?
18:26
You know, I have to say one of the most powerful collaborative tools that I use
18:34
anywhere is Miro. I love that. Right. And it's because I've learned it. Right.
18:38
You can use a
18:39
variety of different tools, but I like Miro. And the reason why is because it
18:42
helps us kind of
18:42
the prototype and visualize concepts and kind of bring them to life for people
18:46
because we're
18:47
very visually oriented. One of my old mentors, Tom Peters, he wrote a great
18:51
book called 40 Years
18:52
ago Called In Search Of Excellence. You said the value of a prototype is that
18:55
someone can point at
18:57
it and say that's not it. Oh, I understand. Right. It solves this ambiguity of
19:00
like you and I just
19:01
sitting around talking. Right. So I really like Miro as a tool that I'm going
19:06
to use all the time
19:07
and I can't live without my slack. Yeah. Right. I mean, slack's like so much
19:12
better than traditional
19:13
messaging products. Slack is so much more collaborative than asynchronous email
19:18
, etc. So I started to
19:21
try to find more and more innovative ways to use slack. I've also tried to
19:26
become much more conscious
19:27
of not reaching out to people on slack during times I should know better. Right
19:33
. So if they
19:33
don't have calendar on, if I'm in a slacksum, but I literally take a 10 second
19:37
journey to Google
19:38
Cal to see where they are. And I've learned that if you know when to get them
19:42
live and once you have
19:43
them live, you can clearly pursue the next decision, slacks a killer
19:48
application right now. I love
19:51
that. I do want to say one thing before you forget because, you know, people
19:54
ask me all the time,
19:55
like, what have you figured out, like, in the last few years that that caused
19:59
you to leave your
19:59
career at Upwork after five successful years and go to G2 and it was this
20:05
reality. I read a book
20:07
several years ago by Daniel Kahneman, a great researcher on statistics and data
20:13
and all that
20:13
type of thing. He just recently passed away. He wrote a book called Noise that
20:17
everyone watching
20:18
should read. He talked about the idea that you should look at data like oil. It
20:23
has grades.
20:25
Right. And crude oil is not really usable by an enterprise. I think that's very
20:29
interesting.
20:30
So what I began to look at a few years ago is the concept of data bias. So data
20:35
is only as good
20:36
as the context in which it was collected. Oh, interesting. Okay. So what I
20:41
figured out, and this
20:42
caused me to be like, I got to go to work for G2, a customer's first party
20:46
voice of the customer
20:47
data is so biased and noisy, it would actually decrease your sales if you
20:51
completely relied on it.
20:52
Because it's the context for that gathering is usually perceived by the
20:57
customers of business
20:58
development conversation. The most biased data in the world is data from a
21:02
buyer before a renewal.
21:03
Like, how are we doing? Well, I'll tell you how we're doing because I want to
21:05
discount,
21:06
I'm getting ready to turn. I've learned that the G2 data, the second party
21:10
verified, well-structured
21:12
data is rocket fuel, especially for training a large language model via buyer
21:20
intent, market
21:21
intelligence, or just informing decision making. So to me, this idea that
21:27
second party data is
21:28
completely better than third party data, obviously, because the context is
21:32
terrible there. We're
21:34
literally making stew a first party data. And that's an insight because when I
21:38
talk to organizations,
21:39
they always went, we never thought of it that way because they're so precious
21:43
about what they
21:43
collect. So that's the reason I joined G2 because they're the leaders in the
21:48
world by far, much more
21:49
than Gartner at capturing high quality, well-structured second party data
21:53
because it's in their DNA from
21:55
GoDard all the way down to the rest of the org. Well, congrats again on being a
22:00
part of that team
22:01
and appreciate you being on the show and looking forward to seeing all the
22:05
great things to come
22:05
out of the research arm. Thank you so much, man. Nice to meet you. Thank you.
22:15
[BLANK_AUDIO]