0:00

Three of the top priorities for us are,

0:02

it's just not an important thing for reach.

0:04

I got a lot of hate for this.

0:06

Why am I not getting the views I

0:08

feel I deserve? Does Instagram shadow ban? I'm

0:11

sure a bunch of people are just

0:13

steaming here to say it. I

0:15

really disappointed myself when you asked

0:17

that question. Yeah, I'm sorry for

0:20

asking. They can be hard. My

0:22

DMs and my comments can be

0:24

pretty dark. Yeah, I can imagine.

0:27

I should double check before I

0:29

answer because we're being recorded. the

0:31

best platform for careers. Hello

0:35

and welcome to today's episode of Build

0:37

Your Tribe. I'm your host Brock Johnson,

0:40

but I'm not gonna bore you with

0:42

a long preamble because you didn't come

0:44

here to listen to me talk. You

0:46

clicked on this episode to hear directly

0:48

from the head of Instagram himself. I

0:50

do want to say that I feel

0:52

so honored and humbled to represent all

0:55

of us content creators and small business

0:57

owners in this conversation. I hope you

0:59

learn something and I hope you walk

1:01

away feeling a little bit more at

1:03

peace with growing on Instagram. You might

1:05

be the most visible head of a

1:07

social media app that I can think

1:10

of. And so at least in terms

1:12

of showing up, putting your neck on

1:14

the line, posting content,

1:16

addressing concerns, what led

1:18

to that decision? I don't know

1:20

if one the most, maybe the... There

1:22

might be more famous ones.

1:24

There's definitely more famous ones.

1:26

But in terms of posting

1:29

on the platform, showing up,

1:31

putting yourself out there. So I used

1:33

to work on Facebook for a long

1:35

time. I worked on the Facebook app

1:37

for almost 10 years. And I

1:39

remember maybe about this might be

1:42

10 years ago now. There's more and

1:44

more, as we're getting really big,

1:46

there's more and more concerns

1:48

and questions and scrutiny. And

1:50

my take was that the

1:53

conversations were clearly going to

1:55

happen about Facebook and social media

1:57

more broadly with or without us.

1:59

would make more sense to be

2:01

part of the conversations than not, even

2:04

if it would be sometimes difficult or

2:06

a lot of work or sometimes really

2:08

unpleasant. I spent a lot of time

2:10

on Twitter back in the day talking

2:13

to journalists and some of our most

2:15

avid critics, but I just thought it

2:17

would be better to be part of

2:20

it than not. So I kind of

2:22

built up that habit or that sort

2:24

of part of my job. And so

2:26

then when I transitioned over to Instagram,

2:29

it just made sense to keep doing

2:31

it. And then over the years I've

2:33

tried to focus more on... Not trying

2:36

to reach everybody. Most people don't care

2:38

what I have to say. But for

2:40

people who are using Instagram to make

2:42

a living or as a creator, I

2:45

wanted to make sure that I was

2:47

speaking to them. So it's a lot

2:49

of fun at times. It can be

2:51

hard. My DEMs and my comments can

2:54

be pretty dark. Yeah, I can imagine.

2:56

But in general, I just think it's

2:58

good to be out there. Speaking of

3:01

those creators and the people who want

3:03

to make money on the platform. What

3:05

are your biggest goals or priorities for

3:07

them? in 2025. I mean there's a

3:10

lot of stuff going on right now.

3:12

I mean we want to be the

3:14

best platform for creators and in some

3:17

ways I think you could make that

3:19

case and other ways we're behind but

3:21

there are a lot of creators on

3:23

Instagram, probably more than any other platform.

3:26

So if we look at how many

3:28

people we think are creators are on

3:30

the platform that looks like a bigger

3:32

number than YouTube channels or most other

3:35

platforms. But you know, you never know

3:37

for sure. And we know that there's

3:39

a lot of competition and we know

3:42

that there's a lot of scrutiny and

3:44

a lot of concerns about our platform

3:46

from the crater community as there are

3:48

about all platforms. And so what we're

3:51

trying to do is better understand the

3:53

broader groups needs and then meet those

3:55

needs. The things that matter most as

3:58

far as we can tell and you

4:00

should tell me, because you probably would

4:02

know better than I, certainly reach. People

4:04

always want more reach and so figuring

4:07

out, but it's not just about how

4:09

much. reach, it's about stability, which is

4:11

very hard. I think any platform that

4:14

guarantees you stable reach is probably lion.

4:16

But certainly there's more and less stable.

4:18

But there are other things too. Like

4:20

I really want us to do more

4:23

as a priority on helping small creators

4:25

break out, which is actually from an

4:27

engineering perspective a bit more difficult. I

4:29

want us to do more to reward

4:32

original creators and identifying who's creating original

4:34

content is also an interesting trick. I

4:36

want us to do to do to

4:39

be. better at trends. It takes still

4:41

too long for things to pop on

4:43

Instagram. So those are maybe three of

4:45

the top priorities for us are valuing

4:48

original content, helping smaller creators break, and

4:50

also having trends just happen quickly. Makes

4:52

sense. In terms of helping smaller creators,

4:55

something a lot of people have noticed

4:57

recently is the pie chart of where

4:59

their reach is coming from. the disconnected

5:01

or the unconnected reach as you call

5:04

it, the reach from non-folars. So I'm

5:06

curious, what motivated that change? Trying to

5:08

grow the pie. So basically, if you

5:10

look at how people share on Instagram,

5:13

creators aside for a second, there are

5:15

more photos and videos, not including text,

5:17

shared in DMs and there are in

5:20

stories every day. And there are way

5:22

more photos and videos shared into stories

5:24

than into feed every day. So yes,

5:26

feed is for average folk to share

5:29

photos or videos that they're really proud

5:31

of, but it's a very small percentage

5:33

of average folk who actually post a

5:36

feed post in a given day, whereas

5:38

for creators it's much higher. So it's

5:40

becoming much more of a public domain,

5:42

whereas stories and DMs are much more

5:45

about where your friends are. Obviously, creators

5:47

are in all of those places in

5:49

different ways. But because the amount of

5:51

content being shared into feed was decreasing

5:54

for a long time... we've needed to

5:56

figure out a way to, that meant

5:58

that the amount of time people were

6:01

spending in feed was decreasing. And so

6:03

that was a negative feedback loop, because

6:05

if there's less time being spent, there's

6:07

less reach to go around, there's less

6:10

people liking comment less, then people will

6:12

post less, and then there's less content

6:14

to see, so then people spend less

6:17

time, and they like and comment less,

6:19

and it was going in a circle.

6:21

So recommendations was really just the only

6:23

viable way to maintain feed as an

6:26

important place. And it is the most

6:28

important place for creators to post. Stories

6:30

are awesome, but your average credit is

6:32

just way more people in feed. And

6:35

so recommendations was a way to, from

6:37

just a pure brass tax business perspective,

6:39

maintain feed and grow feed as an

6:42

important part of Instagram. And we try

6:44

to do so in a way where

6:46

it would help people discover creators and

6:48

other accounts that they would love, but

6:51

might not even knew existed. And so.

6:53

you've seen that play out and you've

6:55

seen and I made it I pushed

6:58

really hard to make sure people could

7:00

see the difference between their connected and

7:02

unconnected reach Because they should know where

7:04

it's coming from and what to optimize

7:07

for might be different or how to

7:09

think about it might be different So

7:11

it was a business need and then

7:13

a user need and try to that's

7:16

what we're always trying to do is

7:18

figure something that's not only good for

7:20

creators, but also good for average folk

7:23

and also good for Instagram and then

7:25

at the overlap is where the good

7:27

stuff happens People are always freaking out

7:29

over algorithm changes and updating. There's this

7:32

new algorithm. You answered recently on one

7:34

of your Friday, Ask Me Any Things,

7:36

which are awesome. You said that you're

7:39

constantly tweaking and working on the algorithm,

7:41

but oftentimes in small ways that we

7:43

might not even notice. Is that a

7:45

daily thing that it's being tweaked and

7:48

worked on? Gotcha. Small changes go out

7:50

every day. Gotcha. It's very rare that

7:52

there's a massive change in how we

7:54

do something. It's much more common that.

7:57

We're just trying to get a little

7:59

bit better one day at a time.

8:01

Gotcha. a huge swing in what's going

8:04

on. It probably isn't there was a

8:06

huge change in how we rank content.

8:08

It's probably more there was a huge

8:10

change in the world. Gotcha. It might

8:13

be Mother's Day. Everybody's posting pictures of

8:15

their mom talking about how much they

8:17

love them. That's actually content for you

8:20

to compete with as a creator. Or

8:22

there's seasonality. Like if you have predominantly

8:24

young people who follow you and schools

8:26

out, you see the time spent on

8:29

social media is gonna grow. if people

8:31

are off work, they're going to spend

8:33

more time on that. So if there's

8:35

a big swing in a given day,

8:38

it's probably more likely the world changed.

8:40

Gotcha. But also, one piece of advice

8:42

I give a lot that I don't

8:45

know I'm good at giving is don't

8:47

overreact to any one post, because something

8:49

might have been weird that day. Look

8:51

for the patterns. So for these types

8:54

of posts, how do they do? And

8:56

then double Donna I want works and

8:58

maybe rethink what doesn't. Talking about messaging

9:01

which I know you believe is a

9:03

central part of the platform and I

9:05

would agree that it's one of the

9:07

real strong suits that Instagram has over

9:10

other platforms There's been a lot of

9:12

mixed messaging no pun intended around DMs

9:14

recently and specifically DM automation So I'd

9:16

love if you could clear the air

9:19

about DM automation through approved partners obviously

9:21

and whether that's allowed on Instagram or

9:23

specifically the whole If you comment this

9:26

word, you'll receive this automated message. Oh,

9:28

interesting. So I think it depends. So

9:30

we certainly, we have an API, it's

9:32

allowed, but sometimes there are some providers

9:35

that are doing shady things. So one

9:37

thing is like, don't ever give anybody

9:39

your password. If you're not asking you

9:42

for your password, that's not probably like

9:44

a trusted partner. But actually most accounts

9:46

that are compromised, so that are self-compromised.

9:48

So that happens a lot. So we

9:51

do want to support this part of

9:53

the ecosystem. We know it's important, but

9:55

make sure you are working with someone

9:57

who's reputable. It's possible sometimes we mistakenly

10:00

catch good actors when we're trying to

10:02

catch bad actors. I'm sure that happens.

10:04

But yes, there are APIs, there are

10:07

tools that are supported and allowed. Gotcha,

10:09

okay. In terms of SEO on Instagram,

10:11

it seems like you guys are really

10:13

working to improve the discoverability of profiles

10:16

and posts. Can you talk more about

10:18

the future of how Instagram might work

10:20

as a search engine? showing us what

10:23

we're looking for and having Instagram results

10:25

pop up when we are searching for

10:27

things. So, I mean, two different sides.

10:29

One, yes, we're trying to do better

10:32

at helping, at having content from creators

10:34

on Instagram surface in search results for

10:36

Google and other searches providers, because I

10:39

just think that's a win-win for everybody.

10:41

So we're starting to try to figure

10:43

out what the opportunities are there. That

10:45

team has been doing. has been lean

10:48

and mean and mighty for a long

10:50

time, but they're doing some good work.

10:52

But we're also starting to invest more

10:54

in search on Instagram, because there's so

10:57

much amazing content, and quite frankly, what

10:59

we call content search, as opposed to

11:01

searching for an account, actually searching for

11:04

some type of content. It's not very

11:06

good on Instagram. And that's not a

11:08

knock on that team. That team has

11:10

been really small for a very long

11:13

time. So we've... strengthen that team recently

11:15

and I'm pretty excited about some of

11:17

the ideas that will hopefully and improvements

11:20

that are going to go out over

11:22

the next well the next few months

11:24

and a few years and it's a

11:26

long road but I do think that

11:29

could also be really good one for

11:31

people who are looking for things you

11:33

can imagine whatever you use Instagram for

11:35

it'd be great to be able to

11:38

find that more easily but also for

11:40

creators because it should allow content to

11:42

resurface so that you don't kind of

11:45

get all the value in those first

11:47

24 or 48 hours. But we'll see.

11:49

Gotcha. One of those little features that

11:51

might be something that that team worked

11:54

on that I noticed recently was at

11:56

the top of the comments section. there

11:58

was a recommended search. Yeah. Can you

12:01

tell me more about that feature and

12:03

what determines what that recommended search is?

12:05

So we're working on it. Actually, the

12:07

latest version is not out yet, but

12:10

what we're looking at is both the

12:12

content, so the real or the carousel

12:14

or the photo. But what we also

12:16

need to look at is the comments

12:19

below. Okay. Because sometimes, and that's not

12:21

out, yeah, but I think that's coming

12:23

out. Sometimes, where the real interesting. context

12:26

is, it's not in the piece of,

12:28

not in the video that someone uploaded,

12:30

but in the context around it, which

12:32

is almost always in the comments. And

12:35

so what we're trying to do is

12:37

surface that more easily. And then you

12:39

can go and find out more. So

12:42

if you saw, I don't know, Walton

12:44

Goggins face and sever, not severance, White

12:46

Lotus this week, there's like a big

12:48

meme. Yeah. You could, maybe you see

12:51

that, maybe you haven't, maybe you don't

12:53

watch the show. and use a picture

12:55

or a real of that. And then

12:57

you'd be like, why does this guy

13:00

look so confused? But in the comments,

13:02

you might find the context between this

13:04

crazy Sam Rockwell sort of monologue in

13:07

this barn in this barn and a

13:09

hotel in Bangkok or whatever it was.

13:11

And then that might be, the prompt

13:13

might be something like, I don't know,

13:16

White Lotus episode seven or whatever it

13:18

was. And so then you would give

13:20

you context. And then if you clicked

13:23

on that, you might get. a whole

13:25

bunch of other content that put it

13:27

in context. Our hope is to sort

13:29

of create a little bit of fun

13:32

and delight by connecting the dots, right?

13:34

Because the real might be funny on

13:36

its own or the meme. The comments

13:38

might be good to read if you've

13:41

got time. You may or may not

13:43

search for it. But if you can

13:45

connect the dots all of a sudden,

13:48

maybe one plus one plus one can

13:50

equal four instead of three. Got you.

13:52

So I know you said the comments

13:54

are coming soon. Right. the caption, the

13:57

spoken audio, the actual like analyzing of

13:59

what is in the... video, all of

14:01

that? All of those things are considered,

14:04

but the technology we're using to actually

14:06

understand those things is going to get

14:08

much better over the next couple months.

14:10

So the way this works for any

14:13

provider is, this is to another platforms

14:15

too, is you don't actually, you basically

14:17

take key frames from the video, so

14:19

you might look at 16 or 128

14:22

or even as many as I think.

14:24

I think YouTube is probably more like

14:26

up to like 512 different frames from

14:29

the video to try to understand that

14:31

would happen. You also will transcribe the

14:33

audio into text and then try to

14:35

use that and extract any understanding from

14:38

that. Comments coming soon if not already

14:40

out. And then from all of that

14:42

you try to come up with a

14:45

search that might be interesting. And we

14:47

try and if we don't come up

14:49

with anything that's interesting, we just don't

14:51

show you. the prompt. So what you'll

14:54

see is those prompts will get, they'll

14:56

happen more often as we get better

14:58

at understanding content and context, and then

15:00

we're more confident in the recommendation. Makes

15:03

sense. I think it's very helpful for

15:05

creators to know that it's everything. It's

15:07

all of those things to be aware

15:10

of. Yeah, and it's going to be

15:12

like, right now it's disproportionately looking at

15:14

the caption relative to the... video relative

15:16

to the audio relative to the comments

15:19

in that order, but that's going to

15:21

change because as we get better understanding

15:23

all the pieces We'll integrate them. So

15:26

definitely just assume we're trying to understand

15:28

the full context Because that's where we're

15:30

going. That's what we're working on. Gotcha.

15:32

Instagram is a beast. There is no

15:35

lie that every single week the algorithm

15:37

changes. There's a million new features. There's

15:39

so much to keep track of so

15:41

much you have to worry about and

15:44

all you're trying to do is grow

15:46

your business, trying to get more customers

15:48

and make more sales on Instagram. That

15:51

is why we created the Insta Club

15:53

Hub to help you learn exactly what

15:55

you need to know, to stay up

15:57

to date with the changes that are

16:00

actually relevant to you as a business

16:02

owner, and to help you grow your

16:04

following and... customer base on Instagram. You

16:07

can check it out, best part, for

16:09

only $7. You can join a two-week

16:11

trial for only $7. That's 50 cents

16:13

per day for two weeks. And all

16:16

you have to do to grab that

16:18

trial offer is go to Insta Club

16:20

hub.com/trial. Again, that link will be down

16:23

in the description. But right now, while

16:25

you're listening, you can just open up

16:27

your safari browser or whatever internet browser

16:29

you use and just go to Insta

16:32

Club hub. trial. You've talked about kind

16:34

of the three metrics that are best

16:36

or that you would recommend focusing on

16:38

the most for increasing reach. What do

16:41

you tell people when they ask you

16:43

what shares? Yeah, because I think shares

16:45

at least algorithmically speaking, of course, they're

16:48

helpful, but it's a human sending opposed

16:50

to another human. So even if the

16:52

algorithm was optimized for saves, a share

16:54

would still be the way to reach

16:57

another human being. That's how I think

16:59

about it. But speaking of this conversation

17:01

of shares, Of course, human to human,

17:04

all shares are not equal. If I

17:06

send a post to my wife, that's

17:08

one additional viewer. If I share it

17:10

to my story, that might be hundreds

17:13

or thousands of extra viewers, algorithmically speaking,

17:15

are all shares created equally or are

17:17

different shares weighted more heavily? Because it's

17:19

true that like if you send a

17:22

message to one person, there's less views

17:24

than if you send it. to your

17:26

stories and maybe thousands of people are

17:29

going to see your story. But they

17:31

might have seen your other stories that

17:33

day anyway. And your wife met another

17:35

open Instagram that day. And when you

17:38

send her a message, her phone's going

17:40

to bing and she's going to open

17:42

it up most likely. Hopefully, I don't

17:45

know. But also because if you just

17:47

think about it from the perspective of

17:49

the person receiving it, like, it's kind

17:51

of more meaningful. If you send me

17:54

something, you're like, I think you'd be

17:56

really interested in this. That is a

17:58

lot more meaningful than. I thought everyone

18:00

might want to know this thing in addition

18:02

to all the other things I thought they

18:05

all should know today. Yeah, yeah. So we

18:07

look more at sentence. Gotcha. And then the

18:09

third metric, if you will, has to

18:11

do with the time, the amount of time

18:13

people are spending with the content. Yeah. I

18:16

want, first of all, clarification.

18:18

It's watch time, so the amount

18:20

of seconds that people are spending

18:22

with that piece of content, not

18:24

necessarily watch rate. Would that be correct?

18:26

Watchrate being the percentage that you

18:29

watch. Yeah, the percentage of completion.

18:31

We look at both. But we try not

18:33

to, we look at both because if you

18:35

watch something all the way to the end,

18:37

that's a good sign. But we definitely

18:39

don't want to just penalize

18:42

things because they're longer. And

18:44

so if you watch 20 seconds

18:46

of a 21-second video and 20

18:48

seconds of a 10-minute video, like,

18:50

it's not like those two things are equal.

18:52

So, two things are equal. So we do

18:55

look at both. We're not trying to

18:57

optimize for time. This is

18:59

a common misconception. I was like, oh,

19:01

these platforms, they just want you

19:03

to spend more time. They're just

19:06

trying to optimize for time. That

19:08

is not quite the case for us.

19:10

I think if you look at Netflix,

19:12

from what I can tell, they optimize

19:14

for time. YouTube, time, tick-talk,

19:16

time, plus some other things

19:18

around originality and breaking content.

19:21

We are a place to be to

19:23

be entertained and attain yourself to

19:25

find stuff that to explore your

19:27

interests But we're also a place

19:29

to connect with your friends And

19:31

if we just optimize for time

19:33

what would happen is you would just

19:35

show longer and longer and more

19:37

and more videos But then for

19:40

every five-minute video you watched that's

19:42

maybe 50 things you didn't see

19:44

from other other people including your

19:46

friends. So we're much more in

19:49

the like short video focus This

19:51

means there's a perfect length for

19:53

video, but one of the reasons why

19:55

the primary flows only allow for up to

19:57

three minutes is like we are a place where

19:59

we will. I want you to easily explore

20:01

a bunch of interests, send a couple

20:03

of things to a friend, maybe talk

20:06

to your friends about them. And short

20:08

videos way more like close to that

20:10

friend connection use case, then long video.

20:12

Because long video becomes much less participatory,

20:14

much more, I'm just gonna sit back

20:16

and watch, right? Which is awesome. It's

20:19

not a bad thing. But it's not

20:21

our thing. I don't think you can

20:23

really have a lean back experience that's

20:25

also when you talk to your friends.

20:27

You kind of have to pick. two

20:30

of the three. You can either be

20:32

long form and short form or short

20:34

form in friends. It's really hard to

20:36

be long form in friends. Makes sense.

20:38

Speaking of hashtags while we're on the

20:41

topic of algorithm and increasing reach, I

20:43

know there's been a lot of changes

20:45

with hashtags over the last few years.

20:47

I believe you can no longer follow

20:49

hashtags. A recent section of hashtags is

20:52

no longer a thing. I know you've

20:54

said a few times now that... I

20:56

got a lot of hate for this.

20:58

Yeah, I know you do. Hash tags

21:00

are no longer a primary way to

21:03

increase your reach on Instagram. They don't

21:05

significantly increase your reach on Instagram. Gotcha.

21:07

Contrary to popular belief. It's just not

21:09

an important thing for reach. So I

21:11

don't know. I can say that all

21:13

I want. Yeah. People may or may

21:16

not believe you. Yeah. I mean, I

21:18

like hashtags. There's something against hashtags. We

21:20

moved away from following hashtags because it

21:22

wasn't really... being used that much and

21:24

just recommendations in general, we're kind of

21:27

handling that. We moved away from the

21:29

recent tab, which I know people really

21:31

like, because it just was a safety

21:33

and a spam. It was getting abused

21:35

by bad actors, I would imagine. Yeah,

21:38

it's much harder to manage the safety

21:40

risks and the spam risks, because you

21:42

can just, if it's just determined, I

21:44

know everyone's just like, just do go

21:46

crowd out, deterministic, recent hashtag, recent feet.

21:49

But for hashtag specifically, if you just

21:51

do that, it's really easy just to

21:53

like. Post a thousand bad things that

21:55

don't quite violate our community guidelines, but

21:57

almost do and just take over that

21:59

page and so it just became like

22:02

now worth it. Whereas if you allow

22:04

us to rank, we can then filter

22:06

out things that are problematic or spammy

22:08

or repetitive or unoriginal and we don't

22:10

have that ability if it's just a

22:13

krano thing and it just was getting

22:15

abused too much. So I get it.

22:17

I'm sure a bunch of people are

22:19

just steaming here to say it, but

22:21

that was I want to be honest.

22:24

I'd rather tell people why and have

22:26

them disagree than just like. pretend it

22:28

was an intentional decision. Is there a

22:30

day where hashtags are removed from Instagram

22:32

completely? I don't think so. I think

22:35

there's probably a couple interesting ways we

22:37

could evolve it. We'll see where search

22:39

goes over the next year or two.

22:41

But there's no plan to just remove

22:43

them. Gotcha. Staying on the topic of

22:45

hashtags. There are some banned hashtags. Hashags

22:48

you can't use. What happens if you

22:50

use one unknowingly on one of your

22:52

posts? Will that affect your account? It's

22:54

not a content violation. Basically what will

22:56

happen is it just won't turn into

22:59

a link. And so we basically, so

23:01

you know, we try to make sure

23:03

we're not encouraging people to discover unsafe

23:05

content, right? So anything related to self-harm,

23:07

for instance, or that's the kind of

23:10

thing that we would just basically block

23:12

the tag. It's not a strike. I

23:14

don't know if it would affect. your

23:16

distribution route, I should double check before

23:18

I answer because we're being recorded. I

23:21

don't think it would, but I would,

23:23

I'd like to double check that, but

23:25

most of the main thing is like,

23:27

we just don't want to make that

23:29

a link, you didn't know, maybe you

23:31

didn't mean it. Sometimes they're actually seem

23:34

like benign names, because a lot of

23:36

people try to work around these systems,

23:38

and obviously there are... certain like self-harm

23:40

related or curse words that might be

23:42

frowned upon, let's call it. Are there

23:45

specific words or phrases that the algorithm

23:47

can detect in our content? and then

23:49

downrank us or our content because of

23:51

it. It's not quite that direct. Okay.

23:53

So there's not like a list of

23:56

words that you're not allowed to use

23:58

and then you're going to get downranked.

24:00

But we do have things like, you

24:02

mean there's two sets of rules that

24:04

are high level. There's what we call

24:07

our community standards, which is like what

24:09

can and cannot be on the site.

24:11

I know you're not allowed to post.

24:13

you're not allowed to celebrate violence, you're

24:15

not allowed to post nudity, you're not

24:17

allowed to post a series of things.

24:20

But then there's also a recommendation guidelines,

24:22

which is content that we feel comfortable

24:24

or not comfortable recommending. And the reason

24:26

we have those is, even though it's

24:28

complicated, is because if you're seeing something

24:31

from someone you follow, then if you

24:33

find it offensive, that's on... the person

24:35

you follow. Yeah. Right. If you just

24:37

post something offensive and I follow you,

24:39

it's like, that's on you. Yeah. And

24:42

it was, and it was your decision

24:44

to follow me. Exactly. So we don't,

24:46

we try not to get between people

24:48

and people they follow. There are, I

24:50

think at this point, zero safety, like

24:53

zero of these like recommendation implementations for

24:55

corrective content. We do not down rank

24:57

content from people you follow because we

24:59

think it might be offensive. Yeah. It's

25:01

not really on that random person, it's

25:03

more on us. So we have more

25:06

responsibility there, so we take more precautions

25:08

there. And so it's possible if you

25:10

write really aggressive content, there might be

25:12

nothing wrong with that. Maybe you are

25:14

advocating for a cause and you're mad

25:17

and you have a right to be.

25:19

That's okay. But it might get cut

25:21

up in, oh, this isn't harassment, but

25:23

it's borderline, it's aggressive, and we won't

25:25

recommend it. So it's possible that. It's

25:28

not like this is specific words, but

25:30

if you are posting something and it's

25:32

getting close to one of our rules,

25:34

it will be less likely to be

25:36

recommended. Nudity is another example. If you

25:39

post something that is like... Borderline nudity

25:41

but not actual nudity. So we will

25:43

not take it down because you know

25:45

exactly where that Line is and you're

25:47

not crossing it. You're going right up

25:49

to that line The chances that you're

25:52

not going to have that content recommended

25:54

because we're going to see that as

25:56

probably Not appropriate for us to recommend

25:58

the people who haven't decided to follow

26:00

you, but we will not get between

26:03

you and your followers. Got you. Okay.

26:05

I do feel like the account status

26:07

area in settings is very helpful being

26:09

able to check and see Is my

26:11

content recommendable? Am I in good standing?

26:14

I'm curious though. Are there ever times

26:16

where people are not recommendable or not

26:18

able to have this disconnected reach and

26:20

yet it doesn't show up in the

26:22

account status? So basically, this is what

26:25

people would say, does Instagram shadow ban?

26:27

Yeah. If there are very few. There

26:29

are a couple real edge cases for

26:31

the most narrowly potential harm you can

26:33

think of, where we might want to

26:35

be. careful not to let someone know

26:38

that we are like looking into something.

26:40

But those are the minority cases. So

26:42

our hope is to put all of

26:44

it in there, everything at the account

26:46

level. It's possible that one piece of

26:49

content might get caught up at your

26:51

account is fine so that won't be

26:53

in there. But at the account level

26:55

we try to have everything be in

26:57

there. The only exceptions I can think

27:00

of are for the darkest of dark,

27:02

which... I hope nobody watching here feels

27:04

like they could possibly fall into that

27:06

category. Yeah, that makes sense. That makes

27:08

sense. And what I also often tell

27:11

people is a ban. All social media

27:13

has their guidelines and they have the

27:15

right to downrank your content or not

27:17

recommend your content, but it's not a

27:19

shadow ban, especially an Instagram's case when

27:21

they're telling you. The word shadow is

27:24

not involved because it's in the light.

27:26

Yeah, and that's the thing that I

27:28

try to explain and I'm always that

27:30

good at it. But I want people

27:32

to know. It's in our interest for

27:35

people to know if you know what

27:37

the rules are and if you know

27:39

what the content is that we're trying

27:41

to reward then you're going to probably

27:43

make more content like that. And that's

27:46

good for you and for us. It's

27:48

not in our interest to like in

27:50

the shadows, like make decisions and now

27:52

look, sometimes we make mistakes. Most of

27:54

those mistakes I hope are reflected in

27:57

account status. You can go there, you

27:59

can see if content is taken down,

28:01

you can appeal, you can see if

28:03

you've got any recommendation issues, you can

28:05

appeal or remove those pieces of content.

28:08

But I want people to know because

28:10

I think it's in our interest for

28:12

people to know. Usually not. Usually, it's

28:14

something else going on. Gotcha. Recently in

28:16

one of your Ask Me Anything Fridays,

28:18

you answered a question that I thought

28:21

made a lot of sense around taking

28:23

a break. And essentially, if you take

28:25

a long break and say eight months

28:27

without any posts on Instagram and then

28:29

you come back, the algorithm doesn't have

28:32

much data that it's been recently trained

28:34

on, and so it's very likely that

28:36

that new post won't perform all that

28:38

great. Yeah. That's true. I don't like

28:40

that that's the case. So I'd love

28:43

us to figure out a way to

28:45

do it. I just want us to

28:47

be honest about the fact that that's

28:49

the case. Do you have any idea?

28:51

Is it like a sliding scale that

28:54

the longer time you spend off, the

28:56

less recent data there is? Or is

28:58

it the once it hits three weeks,

29:00

then it's no longer considered recent data?

29:02

It's between the two? So basically, I

29:04

don't think there's a big difference if

29:07

you take off three months or four

29:09

months, three months or four months or

29:11

four months, I'll double check on exactly.

29:13

how long we retain some of these

29:15

signals for. It's not that we have

29:18

no signals, but like if I follow

29:20

you and you post again, we've got

29:22

some signals on my interest in you,

29:24

but we've got more signals on my

29:26

interest in other people because you've been

29:29

off per a month or two. But

29:31

it kind of, I think you probably

29:33

realize most of that unfortunate loss in

29:35

that first month or two, but it's

29:37

definitely not by design. It's a, it's

29:40

a byproduct of how we build things

29:42

and there isn't. I mean, I'm happy

29:44

about it. It's something I'd like us

29:46

to figure out how to address. I

29:48

think there's probably other things that are

29:50

more important to address, like getting better

29:53

at recency and originality and some of

29:55

the other concepts, but it is... I've

29:57

experienced it, like I've gone off for

29:59

a month, I don't know, I'm on

30:01

a recharge, I had a kid, and

30:04

I came back, I was like, oh

30:06

yeah, I gotta work my way back.

30:08

Yeah, yeah. In terms of consumption and

30:10

like the interaction component, if I'm a

30:12

creator and I create about this one

30:15

niche topic, rock climbing, but I only

30:17

consume content or primarily consume content relating

30:19

to a completely different topic, like cake

30:21

baking. Do those two things affect each

30:23

other at all? Do they interact with

30:26

each other? Or not necessarily? Not necessarily.

30:28

Okay. That'd be kind of cool, actually.

30:30

We don't really use... I'm trying to

30:32

think of even inferred. I can't certainly

30:34

know, not explicitly, but maybe not even

30:36

inferred. We don't use your consumption patterns

30:39

as a signal of how interesting your

30:41

content is for the people who are

30:43

interested in your content. Makes sense. We

30:45

do look at things. Like are there

30:47

groups of people who have shared interests

30:50

and interact a lot, etc. So there

30:52

might be some way it seeps in,

30:54

but if you only post about rock

30:56

climbing and you only like watching bacon

30:58

videos, that's great. Like there's nothing wrong

31:01

with that. What about activity specifically around

31:03

the time of posting? So I have

31:05

a post, I'm gonna make it 10

31:07

a. And so from 945 to 10

31:09

AM, I'm active and I'm responding to

31:12

stories and I'm engaging and then I

31:14

make the post and I'm staying active.

31:16

Does the algorithm reward me for that

31:18

activity? Not directly. Indirectly it could happen,

31:20

right? Because if you're on there a

31:22

bunch and you're messaging people and replying

31:25

to comments or whatever, they might see

31:27

that and they might then tap on

31:29

your profile and then interact with that

31:31

content. And so indirectly it could help.

31:33

But the way the ranking system works

31:36

is it doesn't know that that's why

31:38

that's happening. It's just like, oh, if

31:40

this thing is getting more engagement, it's

31:42

more engaging, it's gonna do better. And

31:44

if you're active in a way that

31:47

is getting people to engage more with

31:49

your content, then yes, that would help.

31:51

But it's not something that we're trying

31:53

to reward. I don't want. people to

31:55

feel like they gotta get on there.

31:58

I schedule posts all the time. I

32:00

schedule mine for early because we're here

32:02

on the West Coast. So a lot

32:04

of my father was in the US

32:06

and I want to get my posts

32:08

out before California's awake and before like

32:11

New York is really into their day.

32:13

Yeah. And I'm not up at 5

32:15

a.m. Well, actually I'm off on the

32:17

5 because I got little kids. But

32:19

I'm not on Instagram at 5 a.m.

32:22

when I scheduled these posts. Yeah, I

32:24

used to be the same way, living

32:26

in California, wanting to post so that

32:28

East Coasters could enjoy it. And this

32:30

was before I scheduled and I'd wake

32:33

up and make the post. Oh, I

32:35

hear that. Oh, my God. There's a

32:37

stressful way to wake up. Before we

32:39

scheduled the post, I say, there's a

32:41

stressful way to wake up. Before we're

32:44

going to shift gears into some new

32:46

features, some things that are upcoming that

32:48

are upcoming. New, exciting features, anything that

32:50

you can share about changes coming to

32:52

live in 2025? No. I had something

32:54

pretty on that. No, yeah, no. I

32:57

really disappointed myself when you asked that

32:59

question. Yeah, I'm sorry for asking. I

33:01

just, with the rise of popularity of

33:03

live on other platforms, live shopping, I

33:05

was wondering if Instagram was working on

33:08

anything like that. Nothing major that I

33:10

can think of. Okay. Another new feature,

33:12

you've mentioned this few times, blends. I'm

33:14

psy about this question. Tell me more

33:16

about blends and kind of how it's

33:19

going to operate supposedly. So the way

33:21

it works is you can start a

33:23

blend with a friend in a thread.

33:25

So if you and I message, there'll

33:27

be a little blend icon on the

33:30

top and we can start a blend

33:32

and it's just a feed of reals

33:34

that you or I or both of

33:36

us might be interested in. Okay. As

33:38

you scroll through it, it'll show you

33:40

who it's for. Oh. So you can

33:43

be like, oh, this one, this one

33:45

is like, like, like, for Adam. You

33:47

have to opt into it because it's

33:49

personal, right? Yeah. You see my- Very

33:51

intimate. Oh yeah. Yeah. Yeah. You gotta

33:54

look into the wild, weird world of

33:56

your friend's interest. Yeah. And then the-

33:58

default action if you're if you sort

34:00

of replies to send the message so

34:02

you're talking about it. And so it's

34:05

pretty fun actually. I don't know if

34:07

it's for everyone. You can also do

34:09

it in group chats too, not just

34:11

one-on-one chats. But look I want us

34:13

to be a place where you connect

34:16

over your interests. We're not the best

34:18

place to sit and just watch TV

34:20

or video. And if I'm running late,

34:22

I'm not gonna DM you on Instagram,

34:24

I'm probably gonna WhatsApp you or hit

34:26

you up on eye message or something.

34:29

But I think we can be the

34:31

best place to connect with the people

34:33

you care about, about the things you

34:35

care about. To me, this sits right

34:37

at the heart of that in a

34:40

really kind of cool way. Yeah. I

34:42

know that the adoption of people using

34:44

the friends feed and the following feed

34:46

might not have been as wild, at

34:48

least from what I hear, people still.

34:51

are unaware that that even exists. So

34:53

is the hope that because this is

34:55

in DM is a place where people

34:57

are spending more of their time already

34:59

that it will be a more heavily

35:02

adopted feature, is that the hope at

35:04

least? To some degree, yeah, but it's

35:06

really not about, I'll be pretty excited

35:08

about it if it's not something that

35:10

drives a lot of incremental engagement for

35:13

the whole platform, but it's like a

35:15

really meaningful thing for a minority of

35:17

people who just really love it. Not

35:19

everything is about growing the pie. Growing

35:21

the pie is good. If there's more

35:23

people spending more time on Instagram, that

35:26

is helpful for us as a business,

35:28

that is helpful for creators, because there's

35:30

more reach to go around. But if

35:32

you only grow by just doing things

35:34

that you can get anywhere else, you

35:37

get the same movie clips and like

35:39

old sports highlights and all this commoditized

35:41

content that you can get anywhere else,

35:43

then it's, yeah, you might grow engagement.

35:45

It's easier to leave Instagram in the

35:48

future. Whereas if there's content that you

35:50

can only find on Instagram or experiences

35:52

you can only have on Instagram, that

35:54

might not make us more money. that

35:56

might not mean that any average crater

35:59

reach goes up, but it might just

36:01

like deepen someone's appreciation for the experience.

36:03

I think that's pretty rat. So yeah,

36:05

this one, I'm actually, there's multiple ways

36:07

this could work out that I'd be

36:09

pretty happy with. Good. Speaking of growing

36:12

the pie and increasing average reach, I

36:14

was reading a study that at least

36:16

according to this one study, average reach

36:18

for reals, carosels, feed images, and stories

36:20

were all significantly down last year compared

36:23

to the year prior. is, I'm curious,

36:25

why is that? Why, and I know

36:27

this is the number one question I

36:29

receive, number one question I'm sure you

36:31

receive is, why am I not getting

36:34

the views I feel I deserve? Yeah.

36:36

I mean, I wish I had a

36:38

great answer that would satisfy people who

36:40

aren't getting as much reaches they feel

36:42

like they deserve. In general, reach is

36:45

going up for some people and down

36:47

for other people. It's much more of

36:49

a blend. The thing that's hard is

36:51

that though, overall, there's more reach to

36:53

go around this year than last year

36:55

than last year than last year, there's.

36:58

Also more people on Instagram and so

37:00

there's more competition right and so like

37:02

if it's 10% more content because either

37:04

Because there's 10% more people and they're

37:06

all posting the same amount a day

37:09

as a year ago It doesn't mean

37:11

there's 10% more things consumed. Yeah, there's

37:13

probably one or two percent more things

37:15

consumed Okay, so that means that there

37:17

can be more competition in practice. There

37:20

are people growing and there are people

37:22

shrinking shrinking and there are people who

37:24

are roughly flat Right? You don't get

37:26

the reach you deserve for long enough?

37:28

You move on. And there's new people

37:31

all the time who either have signed

37:33

up for a while, but now just

37:35

starting to post or they're just signed

37:37

up in the first place. So there's

37:39

also a decent amount of sort of

37:41

churn, for lack of a better word.

37:44

So I want to get someone's content

37:46

to every single person who's really interested

37:48

in it. Because if we can do

37:50

that. The person who sees a video

37:52

or a photo will be happy, the

37:55

creator of the video will be happy,

37:57

and then we'll be happy because both

37:59

of those people... will be happy, don't

38:01

be using Instagram more. Are we perfect

38:03

at ranking? Absolutely not. But we're

38:06

getting better all the time. So it is,

38:08

the thing I try to really explain to

38:10

people is it is in our interest to

38:12

get someone's content to every person

38:14

who's really interested in it. And

38:16

if we're failing to do so, that's

38:19

bad for our business. People think

38:21

that we suppress reach to get people to

38:23

buy ads. It doesn't make any sense.

38:25

It's easy to get people, other people to

38:27

buy it. That wouldn't be a sustainable business

38:29

model. No, no, it would just be no

38:31

one would be there and it just wouldn't

38:34

work. Like we have over 10 million

38:36

advertisers around the world on meta advertising

38:38

and like we've got, that's okay, that's

38:40

a different challenge. This weaknesses and strengths

38:42

there, but we are more than happy

38:44

to grow engagement with one group of

38:47

people and then sell ads to a

38:49

different group of people against that same

38:51

attention. essentially it boils down to supply

38:53

demand. Oftentimes an increase in the

38:55

supply of maybe the number of people

38:57

posting, the frequency of their posting,

38:59

but the demand hasn't necessarily risen

39:01

at the same pace. Is that what

39:04

you would say? Yeah, for a bunch of

39:06

reasons, right? So for a while, the amount

39:08

of time people were spinning on smartphones was

39:10

growing every year. That's roughly flat in a

39:12

lot of countries like the US now. It's

39:14

not like the, for a while, a lot of people,

39:16

like, like, I mean, think about how many

39:18

people, well, well, Maybe this is a little

39:20

bit longer than five years ago, but ten

39:23

years ago. But ten years ago,

39:25

not everybody had a smartphone. Now

39:27

everybody has a smartphone. We're

39:29

buying them less often. Or five years

39:32

ago, maybe not everyone had fast internet

39:34

access. And now most everyone has

39:36

fast internet access. And now most

39:39

everyone has fast internet access. And so

39:41

that's on the demand side, right? Because

39:43

the amount of people who are love

39:45

it if it was. And that would

39:47

be great. So... The thing that I'm more

39:49

focused on is how can I make

39:52

sure that the right values accrue into

39:54

the right people? So I still think

39:56

we're getting better, but I still think

39:58

aggregators as... great as they

40:00

are and they, I'm not anti-aggregator, they get

40:03

disproportionate amount of value relative to what they

40:05

create if you compare them to original content

40:07

creators. Yeah. It doesn't mean they're, they aren't

40:10

valuable, right? If you go around and you

40:12

source all some stuff and you build a

40:14

reach about that, that's cool. But ideally the

40:16

person who came up with the thing and

40:19

made the thing gets more views than the

40:21

person who found your thing. Yeah. And so

40:23

we're trying to do more and more and

40:25

more to find the original content created when

40:28

we can and But there's still room to

40:30

improve there. Last few questions. If Tiktok goes

40:32

away, what's your advice for Tiktok creators migrating

40:35

to Instagram? I mean, if Tiktak goes away,

40:37

that'd be a big deal. I would say,

40:39

look, more and more creators are across multiple

40:41

platforms. I think that's just smart. I try

40:44

to lean into that. I feel like. It's

40:46

good to not have all your eggs in

40:48

one basket. It's my job to make sure

40:51

Instagram is as compelling a platform as possible.

40:53

But it's a great basket. Yeah, exactly. Or

40:55

a great egg. But you can have a

40:57

couple other eggs, that's cool. When you're joining

41:00

Instagram, I think that when there's some things

41:02

that are the same and there's something that

41:04

are different, for any platform, Tik, Instagram, YouTube,

41:06

Facebook, the most important thing is to be

41:09

really clear out about what you're trying to

41:11

accomplish. aligns with that. Sure, reach is great,

41:13

but for what end? Are you trying to

41:16

get brand deals? Are you trying to evangelize

41:18

a cause? Are you trying to sell mugs?

41:20

Like, what are you doing? And then, you've

41:22

got that goal in mind, you then need

41:25

to find the overlap between content that is

41:27

going to help you achieve that goal and

41:29

content that is going to be compelling. Because

41:32

if it achieves that goal and it's not

41:34

compelling, no one's going to see it. And

41:36

if it's compelling and it doesn't achieve that

41:38

goal, why are you doing it? And then

41:41

the third is that it's true to you.

41:43

So that's true of all the platforms. For

41:45

Instagram specifically, if you're new to Instagram or

41:47

you have... and using it as much. I

41:50

think it's important to understand that it's not

41:52

just about short form video. That's a huge

41:54

part of what we do. It's a lot

41:57

of the time, maybe probably more than half

41:59

the time spent on the platform at this

42:01

point. But there's a lot of other things

42:03

too that can kind of be really interesting.

42:06

You can do stories that show behind the

42:08

scenes of how you make your reals. You

42:10

can create a channel and really create a

42:13

deep relationship with your most passionate followers. and

42:15

message them directly. And the coolest stuff happens

42:17

when people connect the dots between these things.

42:19

They tease their video that's gonna come out

42:22

on Sunday and their stories on Friday. Or

42:24

a medicator once that unfolled everybody and then

42:26

refolled a bunch of people every month with

42:28

a different focus. So one month it was

42:31

about women's education. And I was like, okay,

42:33

that's really cool. You should make a video

42:35

on the first of the month explaining why,

42:38

and then point them to that. These are

42:40

the kind of things where it gets really

42:42

fun. So I would, I would, don't feel

42:44

like you have to do everything, but it

42:47

would be good to understand all the pieces

42:49

so you can see if there's anything that

42:51

aligns with your interests. I have to say

42:54

that the diversity of ways you can approach,

42:56

you can. create and engage with the platform.

42:58

Yeah, and you gotta figure out the ones

43:00

that are good for you, because doing all

43:03

of them might be too much. Yeah, that's

43:05

what gets. Yeah. Last question. Besides blends, what

43:07

feature is coming in the pipeline that you're

43:09

most excited about or that can make the

43:12

biggest difference, maybe for creators? I don't know

43:14

if I can say. I don't, I'm already

43:16

like sharing with it. I'm probably gonna get

43:19

a little grief for how much I should.

43:21

I feel like a new thing I can

43:23

tease, but I will say this. I am

43:25

very excited about the long-term potential around partner

43:28

ads. Partner ads allow you to create... content.

43:30

And you can do this deal off. That's

43:32

fine with the brand. And then they can

43:35

ask for the permission to use your creative

43:37

for their own ads and run ads to

43:39

our ad system. And you have to approve

43:41

that as a creator. And you can make

43:44

that deal. However you want to make that

43:46

deal. The reason why I really like it

43:48

might be a little bit academic, but I

43:50

think it matters is that in general, branded

43:53

content. There's no measurement, right? So you don't

43:55

know, as an advertiser, I paid you five

43:57

grand to post this thing that I get

44:00

anything from it. Our ad system has a

44:02

lot of measurement, which means that how much

44:04

you are actually helping the advertiser is going

44:06

to be much better quantified. And that is,

44:09

I think, going to be really good for

44:11

the small creators. Because when an advertiser doesn't

44:13

have measurement, they don't really know. They go

44:16

on their gut. They go with the big

44:18

name that they know or they like. Yeah.

44:20

And so I think, and sorry if you're

44:22

a big creator out there, I think the

44:25

biggest craters are probably overpaid. And the smallest,

44:27

and the smallest, even not just smallest, and

44:29

everybody else, particularly the small creators, are underpaid.

44:32

Absolutely. And this is growing a lot on

44:34

Instagram, the partner had system. And so if

44:36

we can grow that more, I think we

44:38

can help shift more dollars. from the biggest

44:41

creators to the little ones. And I think

44:43

that's pretty cool. Because the whole promise of

44:45

Instagram and really the internet is to help

44:47

you discover your niche interests. So allow anyone

44:50

to be successful. And then there's other things

44:52

we can build on top of that around

44:54

the creator marketplace and helping brands find the

44:57

creators that can help them and help creators

44:59

vet the right brands and make sure they're

45:01

reputable. There's a lot more work to do

45:03

there. Some people know about it. It's grown

45:06

really fast, but we don't hear about it

45:08

a lot. So I think that would be

45:10

the one that I would be the most

45:13

excited about. Awesome. Adam, thank you so much.

45:15

Thanks for having me.

45:17

This me. blast. is a

45:19

I appreciate Yeah, I Cheers. you.

45:22

Cheers. Again, I want I want

45:24

to say you so you

45:26

so much for listening

45:28

today. I I would

45:31

not be here and

45:33

I would not have

45:35

been able to make

45:38

this interview happen happen. if

45:40

it wasn't for supporters

45:42

like you. you. I want

45:44

to ask you to

45:47

subscribe to this podcast

45:49

where we release episodes

45:51

every single week about

45:54

the latest trends, tips,

45:56

and strategies across the

45:58

various social various And

46:00

I do want to

46:03

let you know that

46:05

next week I'm going

46:07

to be posting a

46:09

follow -up to today's follow-up to

46:12

I'm going to break

46:14

down my thoughts and

46:16

my responses to his

46:19

answers to the questions.

46:21

But if this is

46:23

the last time you

46:25

ever hear my voice,

46:28

I want to say

46:30

one last time I

46:32

you so much for

46:35

being here. And as

46:37

always, so much networking. here. Happy Networking.

46:39

So,