WEBVTT

1
00:00:00.000 --> 00:00:08.804
<v Alice>A hardware startup grew to 100K Instagram followers in six months, and spent zero on ads. Luck, or a method?

2
00:00:09.104 --> 00:00:16.520
<v Bob>Its founder says method. And he calls one famous piece of marketing advice the biggest lie. So who's right?

3
00:00:20.820 --> 00:00:30.279
<v Alice>Welcome to Build Wiz AI Show. Big AI ideas, made clear for builders. I'm Alice, and today I'm the one asking the questions.

4
00:00:30.579 --> 00:00:41.826
<v Bob>And I'm Bob! Today it's Samuel Beek's X article on how his hardware startup, Schematik, went from zero to 100K Instagram followers in six months. All organic.

5
00:00:42.126 --> 00:00:47.290
<v Alice>Hold on. A marketing post, on an AI show? Defend yourself.

6
00:00:47.590 --> 00:00:58.152
<v Bob>Easy: his whole case starts with AI. In a world of AI, building products has become easier, so you're competing with way more people than ever. Getting found is the new battleground.

7
00:00:58.452 --> 00:01:01.591
<v Alice>And we're not taking his word for it alone, right?

8
00:01:01.891 --> 00:01:10.537
<v Bob>Never. We'll check him against how Instagram itself says ranking works, and three guests bring data, an investor's view and a skeptic's column.

9
00:01:10.837 --> 00:01:15.316
<v Alice>Start with the why. Why does he think marketing matters more than ever?

10
00:01:15.616 --> 00:01:26.991
<v Bob>Because when everyone can build, the product alone doesn't set you apart. His answer: make sure your customers find you first and like you more. That's why he thinks marketing is actually more important than ever.

11
00:01:27.291 --> 00:01:28.869
<v Alice>Give me a picture for that.

12
00:01:29.169 --> 00:01:41.148
<v Bob>Imagine a street lined with food trucks, all selling great tacos. The truck with the long queue isn't always the best cook; it's the one people found first. Building is the cooking now, and distribution is the queue.

13
00:01:41.448 --> 00:01:44.347
<v Alice>Nice. Does he have proof it pays off?

14
00:01:44.647 --> 00:01:56.215
<v Bob>He points to his pre-seed round, 4.6 million dollars with Lightspeed. He thinks one of the main reasons investors were excited was that they'd figured out distribution by going viral BEFORE even having a product.

15
00:01:56.515 --> 00:02:04.023
<v Alice>Viral before the product even exists. Bold! Lena, you brought an investor's view on exactly this.

16
00:02:04.323 --> 00:02:22.187
<v Lena>I did. Bryan Kim at Andreessen Horowitz, in "In Consumer AI, Momentum Is the Moat", says there's no moat in consumer AI right now, because the landscape is evolving too quickly. His takeaway: whoever builds first, iterates fastest, and distributes best wins.

17
00:02:22.487 --> 00:02:25.386
<v Alice>So distribution is part of the product now?

18
00:02:25.686 --> 00:02:38.129
<v Lena>Close, with one catch he states plainly: traction from distribution only sticks if your product keeps up. It's like a flashy shop window: it pulls people through the door, but the shelves still have to deliver.

19
00:02:38.429 --> 00:02:41.475
<v Alice>Now the spicy part. The biggest lie?

20
00:02:41.775 --> 00:02:54.265
<v Bob>His words. Someone once told him "going viral is not a strategy", and he calls that the BIGGEST lie. In his opinion, the only real way to win on social media is to go viral and grow a lot.

21
00:02:54.565 --> 00:02:55.999
<v Alice>That's a strong opinion.

22
00:02:56.299 --> 00:03:03.936
<v Bob>It is, and we'll test it in a minute. His reason is in a chart caption: one viral day equals 20 times a normal day.

23
00:03:04.236 --> 00:03:05.667
<v Alice>Help me feel that.

24
00:03:05.967 --> 00:03:16.258
<v Bob>Think of a fishing boat: most days you bring in a few fish, then one day you hit a whole school and the hold is full. That's how he sees a viral day, the big catch that outweighs the quiet ones.

25
00:03:16.558 --> 00:03:20.107
<v Alice>But how does a video actually travel that far on Instagram?

26
00:03:20.407 --> 00:03:33.268
<v Bob>Andrew Hutchinson's reading of Mosseri in Social Media Today: to reach new people and grow your following, sends matter most. You want content people forward to others, and Instagram takes that as a signal to show it more widely.

27
00:03:33.568 --> 00:03:35.922
<v Alice>And if you just can't pull that off?

28
00:03:36.222 --> 00:03:44.067
<v Bob>Here he's fair. Option two is to not spend time on socials at all, and he says that's fine too: it really depends on your business.

29
00:03:44.367 --> 00:03:47.139
<v Alice>Okay, the framework. Step one?

30
00:03:47.439 --> 00:03:56.448
<v Bob>Figure out what your audience already wants to watch. People are on Instagram to watch fun videos. They don't go there to learn about features of products they've never heard of.

31
00:03:56.748 --> 00:04:00.322
<v Alice>Wait, doesn't he say exactly where they're watching?

32
00:04:00.622 --> 00:04:04.172
<v Bob>On the toilet or bored. His words, not mine!

33
00:04:04.472 --> 00:04:06.826
<v Alice>So how did he find his formats?

34
00:04:07.126 --> 00:04:17.189
<v Bob>He looked at hardware tags, noted down the formats people love, and thought about how to make similar ones. And the rule they found: don't talk about your product or brand.

35
00:04:17.489 --> 00:04:20.360
<v Alice>Never mention the product? Then why bother?

36
00:04:20.660 --> 00:04:30.723
<v Bob>Think of a street performer: nobody stops for a sales pitch, they stop for the juggling, and then they notice the hat. He says you can mention or show the product subtly, where it makes sense.

37
00:04:31.023 --> 00:04:34.062
<v Alice>Does Instagram's own explanation back that up?

38
00:04:34.362 --> 00:04:46.777
<v Bob>It does. Instagram's explainer, "Instagram Ranking Explained", says Reels are designed to help you discover new things, with an emphasis on entertainment. And most of what you see there comes from accounts you don't follow.

39
00:04:47.077 --> 00:04:49.559
<v Alice>So strangers are the audience.

40
00:04:49.859 --> 00:04:59.748
<v Bob>Exactly. Instagram even surveys people on whether a reel is worth their time or fun. Your feature launch has to pass that test against everything else in the feed.

41
00:05:00.048 --> 00:05:01.431
<v Alice>Steps two and three?

42
00:05:01.731 --> 00:05:12.944
<v Bob>Make tons of videos like the ones that work, and be consistent. He warns that good videos are really hard, and it takes a ton of iterations. Then post every day, multiple times a day if you can.

43
00:05:13.244 --> 00:05:15.732
<v Alice>That sounds like a full-time job.

44
00:05:16.032 --> 00:05:25.236
<v Bob>Early on he spent at least 50 percent of his time on it. It's like learning an instrument: nobody gets good from one heroic weekend, it's the daily practice that does it.

45
00:05:25.536 --> 00:05:28.018
<v Alice>And when a founder can't keep that up?

46
00:05:28.318 --> 00:05:40.552
<v Bob>He didn't step back, he found help. Flo, a creator with more than 250K followers whose hardware videos they were already watching, started making videos with them and later joined full time as Founding Engineer.

47
00:05:40.852 --> 00:05:48.268
<v Alice>Hire the creator, get an engineer too. Two for one! Maya, is the creator route backed by any data?

48
00:05:48.568 --> 00:06:06.716
<v Maya>Partly. Mary Keutelian at Sprout Social, in her roundup of social media video statistics, cites a Sprout report: sponsored influencer content outperforms organic brand content in reach by 92 percent. That's paid partnerships, not a hire, so read it as an echo, not proof.

49
00:06:07.016 --> 00:06:09.370
<v Alice>And the short video format itself?

50
00:06:09.670 --> 00:06:22.496
<v Maya>The same roundup says 52 percent of social users gravitate toward short-form video, under 60 seconds, on Instagram. Think of it as the snack aisle: most shoppers grab something small on the way past.

51
00:06:22.796 --> 00:06:25.835
<v Alice>Step four is numbers. Which ones?

52
00:06:26.135 --> 00:06:34.337
<v Bob>Views, likes, and most importantly watch time. His rule: the higher your watch time, the more viral your video goes.

53
00:06:34.637 --> 00:06:37.838
<v Alice>Is watch time really what Instagram rewards?

54
00:06:38.138 --> 00:06:52.868
<v Bob>Instagram chief Adam Mosseri says it's near the top: as Andrew Hutchinson reports in Social Media Today, the top three ranking signals are watch time, likes and sends. And sends weigh slightly more when you're reaching people who don't follow you yet.

55
00:06:53.168 --> 00:06:54.758
<v Alice>Make that simple for me.

56
00:06:55.058 --> 00:07:05.028
<v Bob>Watch time is like a restaurant noticing which plates come back empty: nobody writes a review, the plate tells you. A reel watched all the way through tells Instagram it was worth the time.

57
00:07:05.328 --> 00:07:07.362
<v Alice>Is that only a Reels thing?

58
00:07:07.662 --> 00:07:19.292
<v Bob>No. Even in the main Feed, Instagram's explainer lists how likely you are to spend a few seconds on a post among the five interactions it looks at most closely. Attention is the currency everywhere.

59
00:07:19.592 --> 00:07:21.737
<v Alice>Did the numbers ever surprise him?

60
00:07:22.037 --> 00:07:33.795
<v Bob>Humbling ones. What he loves is often not what works, and he admits he's super bad at predicting. And videos with his face in them get about 100 times LESS engagement than faceless ones.

61
00:07:34.095 --> 00:07:37.726
<v Alice>Ouch, that stings! Okay, step five?

62
00:07:38.026 --> 00:07:49.912
<v Bob>Experiment, because formats that work will stop working. He says 90 percent of your experiments will fail, and if they don't, they're not experiments. Recently, they found that explainers work really well.

63
00:07:50.212 --> 00:07:51.882
<v Alice>That's a lot of failing.

64
00:07:52.182 --> 00:08:00.956
<v Bob>Think of a chef's specials board: most specials quietly vanish, but the hits earn a permanent spot on the menu. He calls the winners the rotation.

65
00:08:01.256 --> 00:08:04.202
<v Alice>And platforms? Be everywhere at once?

66
00:08:04.502 --> 00:08:20.062
<v Bob>The opposite. With limited resources, put everything on one platform and dominate it; a bit of everything is, he says, probably the worst advice. Then repurpose: their Instagram content does well on YouTube and TikTok, and terribly on X.

67
00:08:20.362 --> 00:08:24.295
<v Alice>Terribly on X, the place he posted this. Love it.

68
00:08:24.595 --> 00:08:35.157
<v Bob>One catch from Instagram's explainer when you repurpose: it aims to make watermarked reels less visible. So upload clean files, not a copy stamped with another app's logo.

69
00:08:35.457 --> 00:08:37.637
<v Alice>When do you add a second platform?

70
00:08:37.937 --> 00:08:46.857
<v Bob>When the engine really works. He puts that at more than 1M views per month, consistently, and only if Instagram keeps killing it.

71
00:08:47.157 --> 00:08:48.478
<v Alice>And paid ads?

72
00:08:48.778 --> 00:09:01.349
<v Bob>Later than you think. In his words, paid is like a drug: you pay your way out of the pain, and it gets addictive. Get good organically first, then run your organic winners as ads for a much better ROI.

73
00:09:01.649 --> 00:09:04.769
<v Alice>Lena, does your investor see paid the same way?

74
00:09:05.069 --> 00:09:17.142
<v Lena>Mostly. Bryan Kim writes that paid acquisition and SEO can still give you a short-term boost in users, but in consumer AI those strategies rarely create lasting retention anymore.

75
00:09:17.442 --> 00:09:20.818
<v Alice>And doesn't Instagram push everyone to buy ads anyway?

76
00:09:21.118 --> 00:09:30.340
<v Bob>Instagram says no: it doesn't suppress content to encourage people to buy ads. Its pitch is growing reach for whoever creates the most engaging content.

77
00:09:30.640 --> 00:09:36.825
<v Alice>Time for pushback. Theo, you've got the market view, and you're skeptical of the viral playbook.

78
00:09:37.125 --> 00:09:53.521
<v Theo>I am: Shubho Sengupta's column in the Hindu BusinessLine, "The cockroach brief: Why virality is not a marketing strategy", draws a line. When virality serves a goal, it's a powerful amplifier. When virality is the goal, it's an expensive and exhausting way to go nowhere.

79
00:09:53.821 --> 00:09:54.964
<v Alice>Got an example?

80
00:09:55.264 --> 00:10:06.558
<v Theo>A protest movement in India he describes: 20 million Instagram followers in 72 hours, and a few hundred people on the street. He puts the conversion rate at about 0.001 percent.

81
00:10:06.858 --> 00:10:09.595
<v Alice>Followers aren't customers. Noted.

82
00:10:09.895 --> 00:10:18.204
<v Theo>His test for any campaign is one question: and then what? Virality without a goal, he writes, is noise with excellent metrics.

83
00:10:18.504 --> 00:10:21.369
<v Alice>Bob, does Beek fail that test?

84
00:10:21.669 --> 00:10:32.881
<v Bob>Not entirely, I think. His step one aims at the people you care about, not at reach for its own sake. Still, my read is that the post never says how many followers became customers.

85
00:10:33.181 --> 00:10:37.370
<v Alice>And can anyone repeat this? We only hear from the ones who made it.

86
00:10:37.670 --> 00:10:47.303
<v Lena>Bryan Kim would agree it's hard. He writes that every startup wants to go viral, but that's becoming harder, and real breakout moments have become more difficult to engineer.

87
00:10:47.603 --> 00:10:50.247
<v Alice>Maya, any caveat on your numbers?

88
00:10:50.547 --> 00:11:00.517
<v Maya>One. They come from Sprout's own reports and surveys, like its 2026 Content Strategy Report, so they describe social users broadly, not your niche.

89
00:11:00.817 --> 00:11:09.173
<v Bob>And remember, every number in the post is the company's own report, the faceless result included. One team's run, not a law.

90
00:11:09.473 --> 00:11:13.615
<v Alice>Time for Builder's Corner. Bob, one thing to try this week?

91
00:11:13.915 --> 00:11:25.720
<v Bob>Spend one evening in your audience's feed, the way he did with hardware tags, and note down the formats that keep coming back. Then make one short video in the best one, about the problem, not your product.

92
00:11:26.020 --> 00:11:28.050
<v Alice>And how do I know if it worked?

93
00:11:28.350 --> 00:11:38.448
<v Bob>Check the numbers Mosseri names in your insights: average watch time, likes per reach, and sends per reach. Keep what holds attention, drop the rest, and try again next week.

94
00:11:38.748 --> 00:11:41.647
<v Alice>Let's land it. What should people remember?

95
00:11:41.947 --> 00:11:51.918
<v Bob>When building gets easier, getting found is the edge. Make what your audience already watches, not product pitches. Post consistently, and let watch time decide.

96
00:11:52.218 --> 00:11:57.973
<v Alice>And from our guests: a viral spike needs a goal behind it, and the product has to keep up.

97
00:11:58.273 --> 00:12:02.032
<v Bob>And save paid ads for later, to boost what already works.

98
00:12:02.332 --> 00:12:07.368
<v Alice>What's still open? For me: how many of those followers ever buy anything?

99
00:12:07.668 --> 00:12:17.024
<v Bob>And mine: does this playbook carry over to software or AI products, without the fun of watching hardware get built? My guess is explainers could be the bridge.

100
00:12:17.324 --> 00:12:22.534
<v Alice>Plus, how long does a winning format last? He only says it will stop working.

101
00:12:22.834 --> 00:12:27.870
<v Bob>That's our look at growing on Instagram without ads. Thanks to Samuel Beek for his X article.

102
00:12:28.170 --> 00:12:40.996
<v Alice>And thanks to Instagram's ranking explainer, Andrew Hutchinson at Social Media Today, Bryan Kim at Andreessen Horowitz, Mary Keutelian at Sprout Social, and Shubho Sengupta in the Hindu BusinessLine.

103
00:12:41.296 --> 00:12:45.948
<v Bob>That's all for today on Build Wiz AI Show. Thanks for listening, and see you next time.
