WEBVTT
1
00:00:03.240 --> 00:00:08.279
Welcome to consumer perspective for agrofood industry high Tech. I'm
2
00:00:08.320 --> 00:00:12.919
Michelle Nagella, behavioral neuroscientist and founder of Neuroscientists, where I
3
00:00:13.000 --> 00:00:17.239
work at the intersection of consumer psychology, sensory science, and
4
00:00:17.320 --> 00:00:21.359
product innovation. And today I'd like to talk about consumer trends,
5
00:00:21.480 --> 00:00:35.320
but more importantly, why understanding consumer behavior requires looking beyond them. Now,
6
00:00:35.359 --> 00:00:38.560
if you've ever downloaded a consumer trend report like I
7
00:00:38.679 --> 00:00:42.920
have many times, you've probably seen the same kinds of themes.
8
00:00:43.439 --> 00:00:50.399
Maybe consumers want healthier foods, consumers want sustainability, consumers want convenience,
9
00:00:50.640 --> 00:00:55.759
they want indulgence. These reports can be very valuable, but
10
00:00:56.039 --> 00:00:59.039
from an R and D perspective, they often leave us
11
00:00:59.119 --> 00:01:04.000
with a really important question, which is now what. Knowing
12
00:01:04.040 --> 00:01:08.159
that consumers want healthier snacks doesn't necessarily tell you what
13
00:01:08.400 --> 00:01:12.400
product to develop. It doesn't tell you why one healthier
14
00:01:12.439 --> 00:01:17.239
product succeeds while another disappears from shelves, and it certainly
15
00:01:17.239 --> 00:01:21.200
doesn't explain the complicated trade offs consumers make every day.
16
00:01:21.640 --> 00:01:26.040
That's because trends describe what people are talking about, but
17
00:01:26.159 --> 00:01:31.319
behavior is about how people make decisions. Those are not
18
00:01:31.400 --> 00:01:34.680
always the same thing. This is one reason I've actually
19
00:01:34.719 --> 00:01:38.120
become really interested in something I call behavioral topic modeling.
20
00:01:38.640 --> 00:01:42.560
Rather than simply organizing conversations into topics, the goal is
21
00:01:42.599 --> 00:01:46.879
to organize them around behavior. So instead of asking what
22
00:01:47.079 --> 00:01:52.400
themes appear most often, we ask questions more like, where
23
00:01:52.480 --> 00:01:58.359
are consumers experiencing frictions with their products? What expectations are
24
00:01:58.400 --> 00:02:03.560
being reinforced, what trade offs are they making, what routines
25
00:02:03.599 --> 00:02:08.439
are products fitting into, and maybe most importantly, what jobs
26
00:02:08.439 --> 00:02:12.680
are consumers actually trying to accomplish with their products. So
27
00:02:12.800 --> 00:02:16.159
let me give you a simple example. Suppose we collect
28
00:02:16.199 --> 00:02:21.080
thousands of online discussions about potato chips. Traditional topic modeling
29
00:02:21.199 --> 00:02:26.800
might identify themes like fun, new flavors or new textures, packaging,
30
00:02:27.560 --> 00:02:33.039
health ideas, maybe price points, or maybe something about protein.
31
00:02:33.560 --> 00:02:36.800
And those are all useful trends that you can see
32
00:02:36.840 --> 00:02:41.800
going on right now, but they still describe categories of conversations,
33
00:02:42.520 --> 00:02:46.319
where when you use and utilize things like AI and
34
00:02:46.360 --> 00:02:50.120
behavioral topic modeling, you can ask a slightly different question,
35
00:02:50.879 --> 00:02:54.800
how are these ideas connected during decision making? So it's
36
00:02:54.840 --> 00:02:58.919
not about counting the dots, counting all the different new flavors,
37
00:02:58.960 --> 00:03:02.840
counting all the different textures, is it accounting them? We're
38
00:03:02.879 --> 00:03:06.960
making a connection of those dots. So, for example, one
39
00:03:07.000 --> 00:03:11.080
behavioral pattern we often see is what I call permission structures.
40
00:03:11.680 --> 00:03:15.800
Consumers aren't simply looking for healthier chips. They're looking for
41
00:03:15.879 --> 00:03:20.680
permission to enjoy chips without guilt. They talk about protein
42
00:03:21.240 --> 00:03:27.080
because it justifies indulgence. They mentioned baked chips, not because
43
00:03:27.120 --> 00:03:30.919
they're necessarily preferring the taste of baked chips, but because
44
00:03:30.960 --> 00:03:35.560
they make the decision feel more responsible. Suddenly, protein isn't
45
00:03:35.599 --> 00:03:39.879
just an ingredient. It's part of a psychological strategy. Another
46
00:03:39.919 --> 00:03:44.240
behavioral pattern centers on trade offs. Consumers often know that
47
00:03:44.360 --> 00:03:48.080
healthier chips may not taste quite as good as the original,
48
00:03:48.599 --> 00:03:51.840
yet they're willing to accept a certain amount of compromise
49
00:03:51.960 --> 00:03:56.639
in certain situations, so packing lunches or maybe snacking at
50
00:03:56.639 --> 00:04:00.879
work or eating after the gym. The decision isn't about
51
00:04:00.879 --> 00:04:04.759
finding the perfect chip, it's about finding the right chip
52
00:04:04.919 --> 00:04:09.400
for that specific moment. That's a very different innovation opportunity.
53
00:04:09.960 --> 00:04:14.439
So instead of asking how to maximize overall liken, we
54
00:04:14.560 --> 00:04:18.959
can begin asking which consumer tensions are we helping resolve.
55
00:04:19.639 --> 00:04:24.120
Behavioral topic modeling also uncovers habits and rituals that are
56
00:04:24.160 --> 00:04:29.360
easy to overlook. Some consumers reserve indulgent chips just for
57
00:04:29.480 --> 00:04:34.120
weekdays right. Others keep individual bags in the pantry because
58
00:04:34.160 --> 00:04:38.759
portion control helps them avoid overeating. Some buy premium chips
59
00:04:38.800 --> 00:04:42.560
when they're entertaining guests, while choosing more value brands for
60
00:04:42.600 --> 00:04:48.360
everyday use. And again, these aren't simply trends, their behavioral contexts,
61
00:04:49.079 --> 00:04:53.160
and context is what really determines whether a product succeeds
62
00:04:53.319 --> 00:04:57.680
or fails. This is where artificial intelligence or AI can
63
00:04:57.759 --> 00:05:03.040
become incredibly useful. Large language models can organize enormous amounts
64
00:05:03.040 --> 00:05:07.560
of consumer language way faster than any research team could manually.
65
00:05:07.959 --> 00:05:11.759
But the real value isn't that AI can find more topics,
66
00:05:11.839 --> 00:05:15.000
because we don't necessarily need more topics. It's that it
67
00:05:15.040 --> 00:05:19.240
allows us to interpret those conversations through a more behavioral lens,
68
00:05:19.800 --> 00:05:24.399
to move beyond counting those mentions and towards understanding the
69
00:05:24.519 --> 00:05:29.639
decisions and motivations that underlie those mentions. Because ultimately, consumers
70
00:05:29.639 --> 00:05:34.439
don't buy products simply because they're healthier, crunchier, or more sustainable.
71
00:05:34.879 --> 00:05:38.120
They buy products because those products solve problems in their
72
00:05:38.199 --> 00:05:44.120
daily lives. Sometimes practical problems, sometimes emotional problems, and often
73
00:05:44.199 --> 00:05:46.959
both at the same time. For food R and D,
74
00:05:47.199 --> 00:05:50.800
that shift in perspective can be really powerful. When we
75
00:05:50.920 --> 00:05:55.480
stop asking only what consumers are saying and start asking
76
00:05:55.720 --> 00:05:58.800
why they're making the choices that they do, we move
77
00:05:58.920 --> 00:06:03.600
from just tracking trends into designing products that better fit
78
00:06:03.839 --> 00:06:08.399
real human behavior. So thanks for joining me for this
79
00:06:08.560 --> 00:06:13.519
episode of consumer Perspective for agrofood Industry High Tech Again.
80
00:06:13.680 --> 00:06:16.959
I'm Michelle Nagella, founder of Nerdoscientists, and I'd love it
81
00:06:17.000 --> 00:06:19.680
if you join me next time as we continue exploring
82
00:06:19.720 --> 00:06:24.199
the behavioral science behind consumer decision making and what it
83
00:06:24.360 --> 00:06:26.319
means for the future of innovation.
1
00:00:03.240 --> 00:00:08.279
Welcome to consumer perspective for agrofood industry high Tech. I'm
2
00:00:08.320 --> 00:00:12.919
Michelle Nagella, behavioral neuroscientist and founder of Neuroscientists, where I
3
00:00:13.000 --> 00:00:17.239
work at the intersection of consumer psychology, sensory science, and
4
00:00:17.320 --> 00:00:21.359
product innovation. And today I'd like to talk about consumer trends,
5
00:00:21.480 --> 00:00:35.320
but more importantly, why understanding consumer behavior requires looking beyond them. Now,
6
00:00:35.359 --> 00:00:38.560
if you've ever downloaded a consumer trend report like I
7
00:00:38.679 --> 00:00:42.920
have many times, you've probably seen the same kinds of themes.
8
00:00:43.439 --> 00:00:50.399
Maybe consumers want healthier foods, consumers want sustainability, consumers want convenience,
9
00:00:50.640 --> 00:00:55.759
they want indulgence. These reports can be very valuable, but
10
00:00:56.039 --> 00:00:59.039
from an R and D perspective, they often leave us
11
00:00:59.119 --> 00:01:04.000
with a really important question, which is now what. Knowing
12
00:01:04.040 --> 00:01:08.159
that consumers want healthier snacks doesn't necessarily tell you what
13
00:01:08.400 --> 00:01:12.400
product to develop. It doesn't tell you why one healthier
14
00:01:12.439 --> 00:01:17.239
product succeeds while another disappears from shelves, and it certainly
15
00:01:17.239 --> 00:01:21.200
doesn't explain the complicated trade offs consumers make every day.
16
00:01:21.640 --> 00:01:26.040
That's because trends describe what people are talking about, but
17
00:01:26.159 --> 00:01:31.319
behavior is about how people make decisions. Those are not
18
00:01:31.400 --> 00:01:34.680
always the same thing. This is one reason I've actually
19
00:01:34.719 --> 00:01:38.120
become really interested in something I call behavioral topic modeling.
20
00:01:38.640 --> 00:01:42.560
Rather than simply organizing conversations into topics, the goal is
21
00:01:42.599 --> 00:01:46.879
to organize them around behavior. So instead of asking what
22
00:01:47.079 --> 00:01:52.400
themes appear most often, we ask questions more like, where
23
00:01:52.480 --> 00:01:58.359
are consumers experiencing frictions with their products? What expectations are
24
00:01:58.400 --> 00:02:03.560
being reinforced, what trade offs are they making, what routines
25
00:02:03.599 --> 00:02:08.439
are products fitting into, and maybe most importantly, what jobs
26
00:02:08.439 --> 00:02:12.680
are consumers actually trying to accomplish with their products. So
27
00:02:12.800 --> 00:02:16.159
let me give you a simple example. Suppose we collect
28
00:02:16.199 --> 00:02:21.080
thousands of online discussions about potato chips. Traditional topic modeling
29
00:02:21.199 --> 00:02:26.800
might identify themes like fun, new flavors or new textures, packaging,
30
00:02:27.560 --> 00:02:33.039
health ideas, maybe price points, or maybe something about protein.
31
00:02:33.560 --> 00:02:36.800
And those are all useful trends that you can see
32
00:02:36.840 --> 00:02:41.800
going on right now, but they still describe categories of conversations,
33
00:02:42.520 --> 00:02:46.319
where when you use and utilize things like AI and
34
00:02:46.360 --> 00:02:50.120
behavioral topic modeling, you can ask a slightly different question,
35
00:02:50.879 --> 00:02:54.800
how are these ideas connected during decision making? So it's
36
00:02:54.840 --> 00:02:58.919
not about counting the dots, counting all the different new flavors,
37
00:02:58.960 --> 00:03:02.840
counting all the different textures, is it accounting them? We're
38
00:03:02.879 --> 00:03:06.960
making a connection of those dots. So, for example, one
39
00:03:07.000 --> 00:03:11.080
behavioral pattern we often see is what I call permission structures.
40
00:03:11.680 --> 00:03:15.800
Consumers aren't simply looking for healthier chips. They're looking for
41
00:03:15.879 --> 00:03:20.680
permission to enjoy chips without guilt. They talk about protein
42
00:03:21.240 --> 00:03:27.080
because it justifies indulgence. They mentioned baked chips, not because
43
00:03:27.120 --> 00:03:30.919
they're necessarily preferring the taste of baked chips, but because
44
00:03:30.960 --> 00:03:35.560
they make the decision feel more responsible. Suddenly, protein isn't
45
00:03:35.599 --> 00:03:39.879
just an ingredient. It's part of a psychological strategy. Another
46
00:03:39.919 --> 00:03:44.240
behavioral pattern centers on trade offs. Consumers often know that
47
00:03:44.360 --> 00:03:48.080
healthier chips may not taste quite as good as the original,
48
00:03:48.599 --> 00:03:51.840
yet they're willing to accept a certain amount of compromise
49
00:03:51.960 --> 00:03:56.639
in certain situations, so packing lunches or maybe snacking at
50
00:03:56.639 --> 00:04:00.879
work or eating after the gym. The decision isn't about
51
00:04:00.879 --> 00:04:04.759
finding the perfect chip, it's about finding the right chip
52
00:04:04.919 --> 00:04:09.400
for that specific moment. That's a very different innovation opportunity.
53
00:04:09.960 --> 00:04:14.439
So instead of asking how to maximize overall liken, we
54
00:04:14.560 --> 00:04:18.959
can begin asking which consumer tensions are we helping resolve.
55
00:04:19.639 --> 00:04:24.120
Behavioral topic modeling also uncovers habits and rituals that are
56
00:04:24.160 --> 00:04:29.360
easy to overlook. Some consumers reserve indulgent chips just for
57
00:04:29.480 --> 00:04:34.120
weekdays right. Others keep individual bags in the pantry because
58
00:04:34.160 --> 00:04:38.759
portion control helps them avoid overeating. Some buy premium chips
59
00:04:38.800 --> 00:04:42.560
when they're entertaining guests, while choosing more value brands for
60
00:04:42.600 --> 00:04:48.360
everyday use. And again, these aren't simply trends, their behavioral contexts,
61
00:04:49.079 --> 00:04:53.160
and context is what really determines whether a product succeeds
62
00:04:53.319 --> 00:04:57.680
or fails. This is where artificial intelligence or AI can
63
00:04:57.759 --> 00:05:03.040
become incredibly useful. Large language models can organize enormous amounts
64
00:05:03.040 --> 00:05:07.560
of consumer language way faster than any research team could manually.
65
00:05:07.959 --> 00:05:11.759
But the real value isn't that AI can find more topics,
66
00:05:11.839 --> 00:05:15.000
because we don't necessarily need more topics. It's that it
67
00:05:15.040 --> 00:05:19.240
allows us to interpret those conversations through a more behavioral lens,
68
00:05:19.800 --> 00:05:24.399
to move beyond counting those mentions and towards understanding the
69
00:05:24.519 --> 00:05:29.639
decisions and motivations that underlie those mentions. Because ultimately, consumers
70
00:05:29.639 --> 00:05:34.439
don't buy products simply because they're healthier, crunchier, or more sustainable.
71
00:05:34.879 --> 00:05:38.120
They buy products because those products solve problems in their
72
00:05:38.199 --> 00:05:44.120
daily lives. Sometimes practical problems, sometimes emotional problems, and often
73
00:05:44.199 --> 00:05:46.959
both at the same time. For food R and D,
74
00:05:47.199 --> 00:05:50.800
that shift in perspective can be really powerful. When we
75
00:05:50.920 --> 00:05:55.480
stop asking only what consumers are saying and start asking
76
00:05:55.720 --> 00:05:58.800
why they're making the choices that they do, we move
77
00:05:58.920 --> 00:06:03.600
from just tracking trends into designing products that better fit
78
00:06:03.839 --> 00:06:08.399
real human behavior. So thanks for joining me for this
79
00:06:08.560 --> 00:06:13.519
episode of consumer Perspective for agrofood Industry High Tech Again.
80
00:06:13.680 --> 00:06:16.959
I'm Michelle Nagella, founder of Nerdoscientists, and I'd love it
81
00:06:17.000 --> 00:06:19.680
if you join me next time as we continue exploring
82
00:06:19.720 --> 00:06:24.199
the behavioral science behind consumer decision making and what it
83
00:06:24.360 --> 00:06:26.319
means for the future of innovation.
