Meeting Title: PPTG - Q2-Q3 Data & AI Priorities Date: 2025-05-06 Meeting participants: Uttam Kumaran, Amber Lin, Bencohen, Daniel Schonfeld, Robert Tseng
WEBVTT
1 00:00:55.300 ⇒ 00:00:56.149 Uttam Kumaran: Hey! Ben!
2 00:00:56.580 ⇒ 00:00:57.650 bencohen: What’s up, man?
3 00:00:57.970 ⇒ 00:00:59.030 Uttam Kumaran: Hey! How are you?
4 00:00:59.740 ⇒ 00:01:02.080 bencohen: I’m good. I think I see a dog behind you.
5 00:01:02.610 ⇒ 00:01:04.972 Uttam Kumaran: Yes, that is this, the big hound.
6 00:01:05.930 ⇒ 00:01:06.780 bencohen: I like.
7 00:01:08.590 ⇒ 00:01:14.110 Uttam Kumaran: He’s the he’s the background. And then people are like, is that a real dog? And then.
8 00:01:14.110 ⇒ 00:01:17.660 bencohen: Can you imagine if that was actually like part of a digital background?
9 00:01:17.946 ⇒ 00:01:23.110 bencohen: I mean, I don’t know. I feel like the gift backgrounds you can do now can have like.
10 00:01:23.370 ⇒ 00:01:29.110 Uttam Kumaran: Movement and stuff. But no, he just he’s just emotional support for me.
11 00:01:29.580 ⇒ 00:01:33.029 bencohen: We all need some. What’s up, Robert? How are you.
12 00:01:33.470 ⇒ 00:01:34.880 Robert Tseng: Hey, Ben? Good to meet you.
13 00:01:34.880 ⇒ 00:01:35.670 bencohen: You, too.
14 00:01:36.707 ⇒ 00:01:41.222 bencohen: let me make sure we were. I was just texting with Dan. But let me make sure that he’s
15 00:01:48.930 ⇒ 00:01:50.190 bencohen: hey, amber
16 00:01:54.750 ⇒ 00:01:55.790 bencohen: one sec.
17 00:01:59.540 ⇒ 00:02:01.360 bencohen: Everybody having a good week.
18 00:02:02.940 ⇒ 00:02:06.680 Uttam Kumaran: Yeah, I feel like it’s just busy as ever on our side.
19 00:02:06.980 ⇒ 00:02:10.300 Uttam Kumaran: otherwise pretty good. I don’t know, Robert, what do you think.
20 00:02:10.850 ⇒ 00:02:18.630 Robert Tseng: Yeah, no, I mean, I I love the starts of the week. I have my most energy at this time. I actually don’t. If you talk to me on Thursdays or Fridays, I.
21 00:02:18.900 ⇒ 00:02:21.293 Uttam Kumaran: My my brain starts to slow down then, so.
22 00:02:23.330 ⇒ 00:02:24.560 bencohen: I appreciate it.
23 00:02:26.880 ⇒ 00:02:32.449 bencohen: Well, we might as well get started. I’m sure Dan will jump on in a second.
24 00:02:32.840 ⇒ 00:02:36.999 bencohen: Brutam. We we already started talking about it. But we.
25 00:02:37.680 ⇒ 00:02:43.469 bencohen: we have had basically pieces of this idea for like 5 years.
26 00:02:43.730 ⇒ 00:02:51.309 bencohen: it’s changed as we went. But basically, technology with AI now is clearly in the right
27 00:02:51.770 ⇒ 00:02:59.064 bencohen: place to address this, but we want to do some kind of a chat Gpt. For pools where
28 00:02:59.840 ⇒ 00:03:03.129 bencohen: I think we train it in 2 ways. We train it with basic
29 00:03:03.610 ⇒ 00:03:07.050 bencohen: low level pool knowledge, basic questions that people
30 00:03:07.180 ⇒ 00:03:18.060 bencohen: pool owners or perspective pool owners might ask. And then the second part would be giving it, like, you know, having it ingest some of our stuff from Zendesk.
31 00:03:18.633 ⇒ 00:03:23.169 bencohen: Obviously, we can go relatively high level with that. But
32 00:03:23.840 ⇒ 00:03:40.369 bencohen: we need to be very aware that we do have on staff highly technical support, and we don’t want to make the experience frustrating. So we’ll have to be very understanding of escalation, like if you try to do what we’d like to do. Now, if we’re like a technical problem
33 00:03:41.810 ⇒ 00:03:58.875 bencohen: on chat gpt. It’ll get you kind of the way there, and then you’ll be frustrated. So we want to make sure that we don’t like try to make this like a solve for everything. It’s more of a gateway to get people on a platform, get their email addresses, get them into our ecosystem and community. And
34 00:03:59.710 ⇒ 00:04:10.819 bencohen: we think it’s a very good top of funnel exercise where people that like, for example, like a lot of people, maintain their own pool. A lot of people don’t. But one thing that is constant is
35 00:04:11.110 ⇒ 00:04:20.770 bencohen: water. Chemistry is like a big thing that people are worried about. Like, is it okay to swim in the pool. Should I wait a day? My pool doesn’t look great. What should I do?
36 00:04:21.610 ⇒ 00:04:33.039 bencohen: It would be very easy to do to just say like, Hey, I don’t know if I could swim my pool like the water looks like shit. And then the agent can basically say.
37 00:04:34.440 ⇒ 00:04:37.010 bencohen: go to the store. Get this test strip.
38 00:04:37.340 ⇒ 00:04:40.650 bencohen: put it in the water, take a picture of it and show it to me, and they’ll say, Oh.
39 00:04:40.960 ⇒ 00:04:44.059 bencohen: based on what you showed me. I think you need more chlorine.
40 00:04:44.250 ⇒ 00:04:46.290 bencohen: I would try by throwing one.
41 00:04:46.440 ⇒ 00:04:47.120 Robert Tseng: One
42 00:04:48.620 ⇒ 00:05:09.160 bencohen: Like, you know, tablet like disk, basically like hockey Puck into your feeder. And if you don’t know how to do that, here’s a here’s like, you know. Take a picture of it, and we’ll assess which feeder you have, and we’ll show you how to do it. So I think that that’s like a free service where we’re just trying to get people to make accounts, use it lightly. And then ultimately.
43 00:05:09.923 ⇒ 00:05:26.080 bencohen: we’d like to get into like actually things with regard to equipment. And then, ultimately, we want to sell people, you know, pay wallets somehow into what I always thought was like the big opportunity for us, which was like some kind of a pool tutor. Because
44 00:05:27.580 ⇒ 00:05:33.889 bencohen: Butam knows this. I don’t know. If you guys know this, a lot of our customers are diy people.
45 00:05:34.090 ⇒ 00:05:37.698 bencohen: That’s probably how we’re differentiated. So
46 00:05:39.490 ⇒ 00:05:43.910 bencohen: the main thing that you have with a Diy is someone that’s like they have some
47 00:05:44.200 ⇒ 00:05:54.210 bencohen: handy knowledge. And they’re probably dangerous enough with tools. But like, they usually need some kind of help. So like they might need a Youtube video, they might need to call somebody.
48 00:05:54.310 ⇒ 00:06:14.609 bencohen: They want to make sure they’re doing things in a safe and accurate way, so that they don’t have problems down the road. It’s very often. And it happens we actually get taken advantage of in our customer service, where people that need a lot of help will basically not hire a pool guy, and they’ll use our customer service team as that support. But I want to start to charge for it.
49 00:06:14.720 ⇒ 00:06:19.180 bencohen: So my thought is like Robert wants to install a new pump.
50 00:06:19.600 ⇒ 00:06:36.949 bencohen: He’s kind of handy. He has some experience with plumbing, some experience with electricity, but he thinks this job might be like just like a step outside of his reach. However, if he could facetime, like his dad, who has a lot of knowledge or the pool tutor through this app.
51 00:06:37.450 ⇒ 00:06:42.399 bencohen: he’d be very good. Just just needs a couple of things, someone over the shoulder, so to speak.
52 00:06:42.990 ⇒ 00:06:46.140 bencohen: That’s what I think is the actual monetization of this app
53 00:06:47.010 ⇒ 00:06:49.929 bencohen: is someone that’s Diy that just needs
54 00:06:50.190 ⇒ 00:06:54.490 bencohen: more support. So I see it as a huge opportunity.
55 00:06:55.750 ⇒ 00:07:01.729 Uttam Kumaran: So I guess a couple of questions there and then. Yeah, I think I want to take this conversation in 2 ways. So one, I want to spend
56 00:07:01.960 ⇒ 00:07:20.970 Uttam Kumaran: hopefully, like, I think, the 1st half talking about the AI opportunity. I just have a couple of questions, and then I could sort of come back to you with a little bit of like at least a scope that we can poke at. So I think I have a couple of things there and then. I want to spend the back half talking a little bit about some things on the data side.
57 00:07:20.970 ⇒ 00:07:21.370 bencohen: Yep.
58 00:07:21.370 ⇒ 00:07:26.440 Uttam Kumaran: So couple of questions I had. You know, we just typically
59 00:07:26.560 ⇒ 00:07:30.050 Uttam Kumaran: ask a few of these, although I’ve been able to answer. I’ll get them out of them. So the
60 00:07:30.180 ⇒ 00:07:37.240 Uttam Kumaran: the main use case is gonna be for just like
61 00:07:38.190 ⇒ 00:07:52.200 Uttam Kumaran: if you were to say it’s gonna be mainly customer support use cases meaning they’re not necessarily like Upsell or like. If you were to prioritize a couple of the objectives of the AI.
62 00:07:52.630 ⇒ 00:07:53.860 Uttam Kumaran: What would they be?
63 00:07:55.040 ⇒ 00:07:57.270 Uttam Kumaran: You’ve you’ve you’ve highlighted the customer support.
64 00:07:57.270 ⇒ 00:08:06.399 bencohen: I think things that are googled, a lot which are very basic, like my pool is green. What do I do, or what is going on. What’s wrong?
65 00:08:06.510 ⇒ 00:08:22.940 bencohen: Those are, I think, the lowest hanging fruit that’s going to be the easiest to onboard people into this experience. For free basic Google has a ton of it in chat, Claude. All of them have. They’ve been trained, they have all that. It’s just basic pool care. Let’s say it like that
66 00:08:23.560 ⇒ 00:08:27.118 bencohen: that’s the easiest one. Because
67 00:08:28.840 ⇒ 00:08:33.880 bencohen: that’s what’s going on out there. And it’s you know how it is
68 00:08:34.120 ⇒ 00:08:43.599 bencohen: Google Reddit Twitter. You don’t have any. It’s not unified. It’s it’s, you know, also, like there’s things that matter like
69 00:08:44.150 ⇒ 00:08:52.150 bencohen: if your pool is green and you’re in Florida versus you’re in New York. Those can mean different things.
70 00:08:52.390 ⇒ 00:09:00.410 bencohen: So like, what happens often is, somebody will go to Google, and they’ll look at a problem that they have. But the conditions are very different. Yeah.
71 00:09:00.870 ⇒ 00:09:03.579 bencohen: I think that if you add a layer of smarts
72 00:09:03.900 ⇒ 00:09:10.266 bencohen: because you can, just probably you can just ask them, where are you? Or maybe the app is just enabling location.
73 00:09:10.790 ⇒ 00:09:16.039 bencohen: you can kind of give them a better answer versus what’s going on now which we we see this
74 00:09:16.540 ⇒ 00:09:18.610 bencohen: when we get inbound things is like.
75 00:09:19.170 ⇒ 00:09:22.839 bencohen: people are just so turned around, you know, at all times.
76 00:09:23.250 ⇒ 00:09:31.129 bencohen: So we have the chance to personalize that. Enough that they get an answer. That’s
77 00:09:31.760 ⇒ 00:09:37.019 bencohen: but we they may get to the truth much quicker, less frustrating, etcetera.
78 00:09:37.430 ⇒ 00:09:38.080 Uttam Kumaran: Yeah.
79 00:09:38.080 ⇒ 00:09:39.149 Daniel Schonfeld: Can you hear me?
80 00:09:39.540 ⇒ 00:09:39.880 Uttam Kumaran: Hey! Dan!
81 00:09:40.160 ⇒ 00:09:41.929 Daniel Schonfeld: Okay, hey, guys, sorry about that.
82 00:09:42.910 ⇒ 00:09:44.690 Uttam Kumaran: No problem. Try to get the alert.
83 00:09:44.690 ⇒ 00:09:45.250 Daniel Schonfeld: It’s
84 00:09:46.090 ⇒ 00:09:50.460 Daniel Schonfeld: I haven’t set up my phone, probably for some reason I don’t get the the meeting alerts.
85 00:09:50.760 ⇒ 00:09:53.210 Daniel Schonfeld: Maybe it’s like on. Do not disturb or something.
86 00:09:54.610 ⇒ 00:09:55.830 Daniel Schonfeld: How far did you get.
87 00:09:56.441 ⇒ 00:10:02.359 Uttam Kumaran: Not that far. I’m just sort of asking some like top level questions about the AI agent. I think the next
88 00:10:02.570 ⇒ 00:10:07.159 Uttam Kumaran: question for me is like this sounds at 1st like a
89 00:10:07.730 ⇒ 00:10:12.089 Uttam Kumaran: basically like driving traffic, and then second, like driving
90 00:10:12.980 ⇒ 00:10:16.710 Uttam Kumaran: revenue. Is that like, if you were to say that the actual, like.
91 00:10:17.100 ⇒ 00:10:18.740 Daniel Schonfeld: Goals, level.
92 00:10:18.740 ⇒ 00:10:30.520 bencohen: Yeah, I think that would be, yeah. I think the goal is you, wanna you wanna knock out and onboard into this. Let’s call it a community to this product, people that are asking low, level things.
93 00:10:30.870 ⇒ 00:10:40.210 bencohen: We’re not out there. There is nothing we are trying to upsell them. We are just trying to get them their answers more quickly. The Google example of my pool is green.
94 00:10:40.650 ⇒ 00:10:48.790 bencohen: and them looking at people in Florida when they’re in Connecticut, and they’re getting bad information, and they just they throw their hands up and give up, and
95 00:10:48.910 ⇒ 00:10:56.289 bencohen: they end up hiring a pool guy, or or just like letting the pool stay green. They just give up. So we want to basically give them an answer that
96 00:10:57.480 ⇒ 00:11:01.250 bencohen: that is for them and is helpful
97 00:11:01.390 ⇒ 00:11:07.540 bencohen: much more quickly. We don’t want them digging all over the Internet, going into forums, going into Reddit.
98 00:11:07.540 ⇒ 00:11:07.890 Uttam Kumaran: Yeah.
99 00:11:08.240 ⇒ 00:11:30.280 bencohen: Hiring pool guys. We want to basically say, your pool is green because you’re not pumping enough. You’re pumping enough water. You have no chlorine. Why don’t you go and get a test strip. Take a picture, send it to the AI. We’ll give you a recommendation of some kind. Basic basic. That’s what people are dreaming of, and they’re trying to shoehorn Google, and other things into it. It just misses
100 00:11:30.410 ⇒ 00:11:43.490 bencohen: now. So yes. So that’s the exercise there for us is collecting all their email addresses. Of course, collating all this data because it can. It can inform a lot of other things that we might want to do. And then
101 00:11:43.880 ⇒ 00:11:50.989 bencohen: my thought is just like the addiction that everyone has with with all these AI agents in their the palm of their hand.
102 00:11:51.150 ⇒ 00:12:02.700 bencohen: as the quality of as they gain trust in this product. They start asking more questions. So finally, they’ve they’ve settled the green thing, but they’re curious about something else I don’t know.
103 00:12:04.390 ⇒ 00:12:11.510 bencohen: they go to. Instead of going to Google or calling a friend, or their dad, or whatever their neighbor, they just go into the app and say, Hey.
104 00:12:11.990 ⇒ 00:12:18.420 bencohen: you know, I just want to move more water I want. I want like my waterfall thing doesn’t look good anymore. I need a more powerful pump.
105 00:12:18.892 ⇒ 00:12:29.347 bencohen: But I want to install it myself. Can you recommend what to buy? And can I even have somebody watch me while I install this, and we up we we upsell them a tutor of some kind.
106 00:12:29.630 ⇒ 00:12:38.589 Daniel Schonfeld: Okay, let’s let’s take a break. Let’s take a breath for a sec. Just the the very, very simple as the 1st iteration is, I just want to prove the thesis. People use an app for chemical
107 00:12:38.710 ⇒ 00:12:39.679 Daniel Schonfeld: for chemicals
108 00:12:40.170 ⇒ 00:12:45.340 Daniel Schonfeld: if we can get them on the lowest hanging fruit, which is just taking care of Ph, things like that.
109 00:12:45.570 ⇒ 00:12:55.129 Daniel Schonfeld: Let’s just keep it stupid, stupid, simple, and just prove a thesis. Somebody will open an app or download it and do the most basic thing they can do. Hold up a test, strip.
110 00:12:55.670 ⇒ 00:12:58.560 Daniel Schonfeld: show it to the app, and the app will tell it how much
111 00:12:58.670 ⇒ 00:13:03.820 Daniel Schonfeld: what they have to do to to keep their pool clean. If we can get people to do the most basic thing.
112 00:13:04.040 ⇒ 00:13:06.100 Daniel Schonfeld: then we can extrapolate from there
113 00:13:06.420 ⇒ 00:13:24.559 Daniel Schonfeld: the possibilities are actually endless if we can figure this out. But let’s let’s figure out how we get an entry app. That’s stupid, simple. I don’t even care if it collects information. We just want to prove the thesis that we can tell someone. Hey, we have an app that tells your chemistry. All you got to do is hold up a stick to it.
114 00:13:25.090 ⇒ 00:13:35.820 Daniel Schonfeld: Done. Then we can get into. We’ll send you a free test kit and get email. Then we can. You know we can build on top of that. But we wanna how do we get something that’s ready in like a week?
115 00:13:35.820 ⇒ 00:13:40.979 Uttam Kumaran: Do you need? Do you need it to be like an Ios app? Can this just be like a web app?
116 00:13:41.200 ⇒ 00:13:47.350 Uttam Kumaran: And and then that’s 1st question. The second question is like, How are you gonna do the initial like distribution.
117 00:13:47.580 ⇒ 00:13:58.249 Daniel Schonfeld: Just through our own media, through our own, through our own. We we’ve got a lot. We’re gonna utilize free media for all of this we have, we have Facebook groups, we have our own email list text messaging.
118 00:13:58.460 ⇒ 00:14:03.480 Daniel Schonfeld: We’ll start there if we can’t get them to do that. Yeah, of course, we can use a web app just to start.
119 00:14:03.770 ⇒ 00:14:06.360 Uttam Kumaran: Let’s just whatever the the and I was just.
120 00:14:06.819 ⇒ 00:14:07.280 Daniel Schonfeld: This.
121 00:14:07.280 ⇒ 00:14:09.180 Uttam Kumaran: That’ll cut the time.
122 00:14:09.440 ⇒ 00:14:11.820 Uttam Kumaran: And so I guess that’s a good question
123 00:14:12.020 ⇒ 00:14:14.710 Uttam Kumaran: you’re aiming for, like, okay, something we can
124 00:14:15.110 ⇒ 00:14:17.610 Uttam Kumaran: touch and feel in like 2 weeks.
125 00:14:18.700 ⇒ 00:14:19.020 bencohen: And.
126 00:14:19.020 ⇒ 00:14:23.639 Daniel Schonfeld: At most at the most stupid simple like. Right now, I just use chat. Gpt, I could do it today.
127 00:14:23.930 ⇒ 00:14:24.630 Uttam Kumaran: Yeah, yeah.
128 00:14:24.630 ⇒ 00:14:28.569 Daniel Schonfeld: And I can just point it at it, and it’ll tell me so. All we got to do is wrap it.
129 00:14:29.000 ⇒ 00:14:51.610 Daniel Schonfeld: plug it into one of one of the the models and start there literally don’t very little development. We just want to say, Hey, we started pool AI, and we’re it could tell you what your chemistry is. I don’t care if people like. Oh, it’s just chat, Gpt. Just want to see people if it’s accurate, if it makes sense and figure out if people would use this, give it to a hundred people.
130 00:14:51.810 ⇒ 00:15:00.460 Daniel Schonfeld: and I have friends who will use it right away. That use chat. Gpt, I’m gonna say, is, the chat gpt better than this. Maybe we missed something. I don’t want to build all this shit. And
131 00:15:00.460 ⇒ 00:15:04.249 Daniel Schonfeld: yeah, yes, one stupid thing. We realized it was all worthless.
132 00:15:04.590 ⇒ 00:15:08.790 Uttam Kumaran: So okay, we have, like a 2 week, like proof of concept.
133 00:15:09.030 ⇒ 00:15:26.589 Uttam Kumaran: we’re mainly looking for like anecdotal ish feedback. But we’ll see data on if they’re using it. Yeah, doing it as a web app that that should be fine. But the other stuff is we. We just need as much documentation from you on this topic as.
134 00:15:27.200 ⇒ 00:15:31.490 Uttam Kumaran: Like humanly possible. I mean, I’m I’m sure a lot of it exists within chat. Gbt, but
135 00:15:31.650 ⇒ 00:15:33.890 Uttam Kumaran: I don’t want like the generic.
136 00:15:34.200 ⇒ 00:15:35.099 bencohen: Yeah, no, we need to go.
137 00:15:35.100 ⇒ 00:15:45.769 Uttam Kumaran: Whatever the if you if you know either if you guys have the source of truth, or if the source of truth exists somewhere publicly that we can go like use as a reference, like if there’s a pool kind of.
138 00:15:46.349 ⇒ 00:15:46.929 bencohen: I.
139 00:15:46.930 ⇒ 00:15:51.550 Daniel Schonfeld: I would just go. Sorry, Ben, I would. When I did this test myself
140 00:15:52.110 ⇒ 00:16:01.270 Daniel Schonfeld: and chat, gpt, you hit deep research. It goes to trouble. Free pool it goes. It’s actually scours the Internet for all the right answers, and then provides the source.
141 00:16:01.560 ⇒ 00:16:02.420 Daniel Schonfeld: So
142 00:16:02.530 ⇒ 00:16:10.269 Daniel Schonfeld: I’m being serious when I say this. So like. I don’t care if they know it’s chat. Gpt, I don’t. All I care about is if
143 00:16:10.480 ⇒ 00:16:19.449 Daniel Schonfeld: somebody tells me and I give this to my friend who actually does this. It’s a housewife in in Westport that does everything through chat. Gpt, I’m just gonna say.
144 00:16:19.810 ⇒ 00:16:33.300 Daniel Schonfeld: does this work the exact same way? Is there anything that I’m missing and just wrapping it with a with a nice web page and blinking it, or plugging it into chat? Gpt! Did we miss something even on that basic step.
145 00:16:33.510 ⇒ 00:16:39.719 Daniel Schonfeld: and she says, No, this is great. I mean, I wouldn’t use it because I just use chat gpt, and it’s an app, but that’s an answer I can accept.
146 00:16:39.820 ⇒ 00:16:47.839 Daniel Schonfeld: But if she says Oh, it takes a fucking while to load, if for some reason not getting the same information. When I put in the chat like, that’s where I want to find the issues.
147 00:16:49.160 ⇒ 00:16:51.339 Daniel Schonfeld: I just want her to say, yeah, it’s the same exact thing
148 00:16:51.610 ⇒ 00:17:07.510 Daniel Schonfeld: great. And then we can teach it things and say, Okay, we just all we have to do is make it figure out what’s what’s lacking in chat gpt because it’s just taking basic information from the web. Then how do we add personalized intelligence? How do we make it more personalized? How do we go deeper.
149 00:17:07.849 ⇒ 00:17:15.089 Daniel Schonfeld: you know, and and diagnose bigger issues. I just want to start with the most basic dumb thing without too many variables.
150 00:17:15.099 ⇒ 00:17:15.419 Uttam Kumaran: Yes.
151 00:17:15.420 ⇒ 00:17:19.150 Daniel Schonfeld: Somebody says that, yeah, this is, I just want it to be one for one with with
152 00:17:19.540 ⇒ 00:17:24.670 Daniel Schonfeld: with with Openai, or or whatever you want to choose, but I think it’s probably the best one.
153 00:17:25.310 ⇒ 00:17:39.459 Uttam Kumaran: I think, even with like, if you’re if you can do even just like one more tightening like, okay, I hear you on going to trouble free pools. And we can get, we can basically list out like the core sources I actually want to like, try to limit it. Otherwise
154 00:17:39.590 ⇒ 00:17:48.600 Uttam Kumaran: it may, it will at some point make something up, especially as you ask it more specific questions. So we want to understand what we know and don’t know. For example, if someone’s like
155 00:17:48.800 ⇒ 00:17:55.839 Uttam Kumaran: takes a picture of a model number and asks a question, and it doesn’t have the information. It will, as you know, probably try to make something up.
156 00:17:56.370 ⇒ 00:18:00.589 Uttam Kumaran: We wanna flag those as like opportunities for us. And maybe it will say
157 00:18:00.740 ⇒ 00:18:18.520 Uttam Kumaran: that I don’t know right. But that’s fine for this proof of concept. That’s where I almost want to go tighter. And so we do. So to give you context, we do these like proof of concepts for a bunch of clients. Usually it like I said, 2 weeks. Usually it takes like a few days. But I I wanna buy us a little bit of time.
158 00:18:18.971 ⇒ 00:18:34.059 Uttam Kumaran: So I think that that makes sense. The other piece I need is sort of like, if you could just narrate a couple of categories of questions, and we’ll go fill out the rest with AI. But if you just give me a couple of like the common question types
159 00:18:34.490 ⇒ 00:18:38.569 Uttam Kumaran: or categories like, however, you see that you can just rattle them off.
160 00:18:39.300 ⇒ 00:18:41.680 Uttam Kumaran: and that’s probably what we need for now.
161 00:18:42.200 ⇒ 00:18:45.290 Daniel Schonfeld: I would. I would still rely on
162 00:18:45.750 ⇒ 00:19:02.080 Daniel Schonfeld: on AI to find out what those questions are like. I would rather you ask AI is going to do a much better job if you say, what are the most common questions people ask about pool chemistry online? AI is going to do a better job than me, Cody and Ben combined, and it’ll give it to you in 4 min.
163 00:19:02.080 ⇒ 00:19:07.119 bencohen: I mean I’ll I mean, I just prompted. I’ll tell. I mean, I just prompted it. So you’ve got
164 00:19:07.310 ⇒ 00:19:18.280 bencohen: here. I’m just. I’ll send you this, but like, how often should I test my pool water? How do I balance my Ph. Alkalinity and chlorine? What’s the ideal chlorine level for my pool.
165 00:19:18.390 ⇒ 00:19:27.850 bencohen: Why is my pool cloudy or green? How do I get rid of algae? And then, you know equipment? How long should I run my pump for each day. What size, pump, or filter do I need for my pool? How do I?
166 00:19:28.050 ⇒ 00:19:30.439 bencohen: When should I backwash my filter.
167 00:19:30.820 ⇒ 00:19:33.810 Uttam Kumaran: I guess my question for you. Is there anything missing from that list?
168 00:19:34.160 ⇒ 00:19:40.870 bencohen: This is no, these are the hits. I mean, this is the level one, you know.
169 00:19:41.120 ⇒ 00:19:46.490 Daniel Schonfeld: Yeah, start with, start with Basic. And I’ll you like I’ll be the 1st tester. I’ll just go out to my pool. I’ll put a test strip in
170 00:19:46.840 ⇒ 00:20:11.949 Daniel Schonfeld: and start there. I’ll have my friend do it. We’ll start from there. If that works well on the most basic 2, 3 questions with a web page. Let’s move on. Then we’ll add questions. I want to start this thing the right way, because more often than not we do builds, and we we build stuff, and there’s 10 new variables, and we don’t know which one affected the outcome. Just want to start with building blocks from day 1, 1 to 3 questions, stupid, simple.
171 00:20:12.370 ⇒ 00:20:17.669 Daniel Schonfeld: using a back end that’s reliable, and does everything work even with those 3 simple questions?
172 00:20:18.440 ⇒ 00:20:28.490 Daniel Schonfeld: And is it the same experience, if not better or worse than just going into Chat Gpt. And the answer is, probably gonna be chat, gpt is easier. It’s an app. You use it every day. You’re familiar with it.
173 00:20:29.055 ⇒ 00:20:41.269 Daniel Schonfeld: But we need. We need to make it look feel very similar to that. And then we’ll grow from there based on feedback. Everything will be data driven. And the good thing is is the audience we have for free
174 00:20:42.038 ⇒ 00:20:48.400 Daniel Schonfeld: the knowledge base we have internally versus both personalized and the
175 00:20:48.590 ⇒ 00:20:51.530 Daniel Schonfeld: the generic binary data is out there
176 00:20:51.890 ⇒ 00:20:57.240 Daniel Schonfeld: on flow rates. So the if you think about the possibilities, they’re actually endless.
177 00:20:57.350 ⇒ 00:21:09.289 Daniel Schonfeld: Okay for diagnosing issues. We can tap into our Zendesk tickets, we can tap into the entire Internet of reviews. And what people complain about with pumps, etc. We can do comparison shopping.
178 00:21:09.450 ⇒ 00:21:36.829 Daniel Schonfeld: We can diagnose problems in real time and also refer service people. If we can’t diagnose the problem. It’s a free lead tool. I mean, I I it’s it’s incredible what the opportunity is. We just got to start before we go crazy, because there’s so many directions to go with. Just the most basic thing. Could we wrap, basically chat, gpt, and get someone to use it and still have a a decent experience with just the most basic thing.
179 00:21:37.510 ⇒ 00:21:45.679 Uttam Kumaran: Okay, cool. So yeah, I think we probably like, if we can put a time box to next week. But I think we can probably have something for you.
180 00:21:45.950 ⇒ 00:21:51.290 Uttam Kumaran: either by Friday or or earlier. We do. Yeah.
181 00:21:51.720 ⇒ 00:21:53.709 Daniel Schonfeld: Say, keep this super cheap.
182 00:21:54.160 ⇒ 00:21:54.890 Uttam Kumaran: Yeah, yeah.
183 00:21:54.890 ⇒ 00:22:05.110 Daniel Schonfeld: We want to spend money on the big development stuff as we prove mini thesis out like, I don’t like use AI to build the web page like, don’t use developer like, try to keep this super cheap.
184 00:22:05.110 ⇒ 00:22:09.419 Uttam Kumaran: No, that’s what that’s what we do. I’m I’m saying 2 weeks, just because I have to.
185 00:22:09.700 ⇒ 00:22:25.059 Uttam Kumaran: I can’t promise things in like 2 days. But it won’t take. It’s not. Gonna take that long. We typically turn around these sorts of proof of concepts in like 2 days. Basically, for example, we’re speaking with a lot of law firms where we’re doing like document analysis, things like that.
186 00:22:25.060 ⇒ 00:22:25.820 Daniel Schonfeld: Yeah, yeah.
187 00:22:25.820 ⇒ 00:22:28.530 Uttam Kumaran: As part of our sales cycle. We basically put
188 00:22:28.640 ⇒ 00:22:31.769 Uttam Kumaran: in like a day. We’ll we’ll get like something together that basically.
189 00:22:31.930 ⇒ 00:22:35.060 Uttam Kumaran: it’s like, probably halfway or more towards what they want.
190 00:22:35.524 ⇒ 00:22:44.350 Uttam Kumaran: And we use AI to build the whole thing and so I hear you on just making it pretty basic. I think the couple of like things that are probably
191 00:22:46.180 ⇒ 00:22:55.829 Uttam Kumaran: I mean, apart from like having better prompting, having it just focus on pool stuff like that I think we’ll we’ll try to enable, if we can, the image, or at least I’ll tell you
192 00:22:55.990 ⇒ 00:22:57.530 Uttam Kumaran: how long it’s gonna take
193 00:22:57.730 ⇒ 00:23:17.250 Uttam Kumaran: like so that you can take a picture and add it, but I think it should be fine. Apart from that, anything around like Ui and stuff like that, we’ll we’ll just take a stab at we can that all that stuff doesn’t really matter. You can make a decision later. It’ll just be a lot of the reason why I want to limit the knowledge is the better the limit is the better. The answers were. So
194 00:23:17.250 ⇒ 00:23:17.680 Uttam Kumaran: yep.
195 00:23:17.680 ⇒ 00:23:29.810 Uttam Kumaran: it’s good. We’re focusing on chemistry. We’ll get a lot of test data on what people are. Answer, asking what people are, what the answers need to be and like, we’ll go from there. So I about this.
196 00:23:29.810 ⇒ 00:23:36.340 Daniel Schonfeld: Yeah, and don’t worry about if it’s janky. If the results aren’t great, I just want to get something up and running that I can test internally. This is not going out to.
197 00:23:36.340 ⇒ 00:23:40.060 Uttam Kumaran: Do you care if it’s brand? Do you want it to be branded like or.
198 00:23:40.060 ⇒ 00:23:44.870 Daniel Schonfeld: You can. You can, just, you know, make up something like.
199 00:23:45.151 ⇒ 00:23:47.400 bencohen: Give it a blue. Give it a blue.
200 00:23:47.710 ⇒ 00:23:51.420 bencohen: you know. Give it a blue feeling. But this is really just
201 00:23:52.040 ⇒ 00:23:55.559 bencohen: 5 people internally are gonna play with it and say, Okay.
202 00:23:55.720 ⇒ 00:23:59.200 bencohen: we like it. Let’s go to the next step, you know. Spend another
203 00:23:59.310 ⇒ 00:24:04.239 bencohen: 7 to 10 days on it, and then we’ll put it out to a smaller, a tight group, and then.
204 00:24:04.460 ⇒ 00:24:08.420 bencohen: if the feedback’s good there, we’ll continue onwards.
205 00:24:09.620 ⇒ 00:24:36.250 Daniel Schonfeld: Very tight lip also. So my my goal with this is within probably 2, 3 months, is to shift budget away from Google Facebook, which we spent a couple $100,000 a month, plus whatever the number is in peak and start siphoning off that budget into this. So I I think this could be a great customer. Acquisition, tool, lead, magnet, subscription service down the line. But to me the most value right away is new customer acquisition.
206 00:24:36.849 ⇒ 00:24:42.529 Daniel Schonfeld: And by getting people to use it if it’s free, if it’s easy, if it’s simple and it helps.
207 00:24:42.930 ⇒ 00:24:55.140 Daniel Schonfeld: You know, those are just new customers potentially to to pull parts to go. So if we look at it that way, it’s just shifting ad add budget towards more development and and bettering the
208 00:24:55.690 ⇒ 00:25:07.360 Daniel Schonfeld: You know the the customer acquisition side of our funnel. And so I think this is just day. One is just get something so stupid. Basic, easy questions layups and keep having small wins
209 00:25:07.550 ⇒ 00:25:31.369 Daniel Schonfeld: and then using customer feedback to to grow it, not just assuming things and building things for the sake of building them. But all data driven. The big thing also is just make sure there’s there’s good analytics plugged into it, even if it’s basic Ga or whatever. So we can at least see, you know, where people are spending time on this website. How much time it’s taking to load things like that. So we can at least troubleshoot any errors in the 100 people.
210 00:25:31.720 ⇒ 00:25:32.330 Uttam Kumaran: Okay.
211 00:25:32.490 ⇒ 00:25:44.660 Uttam Kumaran: so I’ll send a couple of things over. So, Ben, I actually sent this in the chat. These are like couple of demos that we’ve done for some clients recently. This is sort of the fidelity of something we could produce in like a few days.
212 00:25:45.331 ⇒ 00:25:51.509 Uttam Kumaran: Again, we would just add the we basically add the image and brand it a little bit more. But this isn’t anything
213 00:25:52.190 ⇒ 00:25:59.430 Uttam Kumaran: like crazy for us. So I hear you, I think. Let’s let us take a stab on amber on our side. We’ll basically put together like what
214 00:25:59.820 ⇒ 00:26:03.739 Uttam Kumaran: 2 or 3 tickets look like for this. We’ll ask Miguel. And then we’ll
215 00:26:04.300 ⇒ 00:26:14.660 Uttam Kumaran: basically just send like us, like, basically like, okay, if we were to spend 2 days on this, here’s what we’re gonna achieve, feature wise, and then just try to go for it. So we’ll turn that around.
216 00:26:14.660 ⇒ 00:26:15.810 Amber Lin: Sounds good.
217 00:26:16.600 ⇒ 00:26:17.219 Uttam Kumaran: Thanks, man.
218 00:26:17.450 ⇒ 00:26:18.030 Daniel Schonfeld: That’s great!
219 00:26:18.030 ⇒ 00:26:31.319 Uttam Kumaran: Yeah, I I think you’ll, yeah, I, you’re right. So we we’re using AI to build all this stuff. Yeah, we’re on sort of overdrive internally, using AI as well. It just helps to like very much narrow the scope
220 00:26:31.835 ⇒ 00:26:36.390 Uttam Kumaran: on the on what the AI is doing. Otherwise. Well, yeah.
221 00:26:36.390 ⇒ 00:26:40.503 Uttam Kumaran: quality will go will just go down. So those are the those are the questions. So I I think,
222 00:26:40.900 ⇒ 00:27:02.080 Uttam Kumaran: I wanted to. I know we don’t have a bunch of time left, but we I wanted to spend a little bit of time on the data side. So I spoke with Ben this week, and sort of I know we got a little bit of feedback on how you know. We fixed a bunch of things recently, but I think the dashboards were really out of use for a while, and I know when we started the engagement we were much more active on our side on especially on the analysis piece.
223 00:27:02.640 ⇒ 00:27:14.000 Uttam Kumaran: Got a we got a bunch of quick wins. I think we started focusing more on accuracy. And then, you know, we started doing a couple of other things for the company. But I think we want to go back to basically
224 00:27:14.490 ⇒ 00:27:19.870 Uttam Kumaran: basically sizing the questions by, like, the opportunity for y’all versus
225 00:27:20.440 ⇒ 00:27:37.299 Uttam Kumaran: focusing on things that that aren’t important. We did a couple of. We prepared a couple of notes there. Maybe I’ll let Robert just take that on and talk through like how we can. Actually, we have a couple of opportunities, including the pool, variable pool pump profitability.
226 00:27:37.776 ⇒ 00:27:43.053 Uttam Kumaran: That basically like should hopefully go directly to driving top line, revenue, or profit.
227 00:27:44.160 ⇒ 00:27:46.159 bencohen: Yeah, let’s let’s hear it. That’s great.
228 00:27:46.160 ⇒ 00:28:13.529 Robert Tseng: Yeah, so I guess I’ll do a quick intro, just because I don’t think I’ve actually met Ben and Dan for this call so. I’m really Tom’s business partner. Just you haven’t heard from me, because, you know, Tom’s kind of been doing kind of. He’s kind of been with you guys for a while. And you know, seems like you guys had the relationship. I think, bringing me in at this point like my background, I, I ran a data insights team at Ruggable, which is like a P. Backs Dtc rug rug brand based in la
229 00:28:13.840 ⇒ 00:28:33.289 Robert Tseng: and so I think, based on kind of what he was telling me about what you guys are looking for. And you know now that the data is all clean, we have a lot of data like, how do we actually turn that into insights? I think it’s really the onus is on us to kind of start to size kind of like which I mentioned the opportunities for the type of work that we could be doing for you.
230 00:28:33.693 ⇒ 00:28:57.409 Robert Tseng: And so I think I kind of just took like a 1st pass, and I’ll share this more broadly as well. I don’t. Wanna I know we only have a couple of minutes. So yeah, you know, I think, really like, there’s a few exercises that I think would be really helpful for us to kind of get to, so that we could really start to turn turn on this like insights engine. I guess one is really from like a market research perspective, like really understanding
231 00:28:57.410 ⇒ 00:29:18.519 Robert Tseng: how, how much like, how? Where are you of the total market? Like kind of your your current penetration of the market right now? And then, you know, once we able to do that, even though you know, we may not hit the number like on the nail. It’s really just like a a way for us to start to quantify the. You know, the the opportunity size for the questions that we would ask.
232 00:29:18.610 ⇒ 00:29:26.249 Robert Tseng: And then, obviously, I’ve I’ve looked through your product portfolio. I’ve been setting the data from the past few days and understanding like, okay, like.
233 00:29:26.830 ⇒ 00:29:35.329 Robert Tseng: you know, there’s a few things one is like as you’re launching new skews like, what’s the velocity that you’re moving at like? How long does it take for you to get from time to listing
234 00:29:35.908 ⇒ 00:29:37.859 Robert Tseng: and like, how do we really help?
235 00:29:38.080 ⇒ 00:30:04.800 Robert Tseng: I kind of speed up that product development process, make sure that you’re launching products into the right categories with with high growth. Potential. You know. So these are like big consulting, like questions. I suppose, that are, I think, helpful from a strategic perspective, so that we can align like the analytics that we’re doing more closely to your product development roadmap and how you think about business. I mean, I have a list of questions in each of these sections that I’ve outlined. You know, one is.
236 00:30:04.800 ⇒ 00:30:15.690 bencohen: Here’s what I would say. I can. I can chat with you, and we can do this. This is very high level stuff. We we need like tactical help, like I. I talked to Tom on Thursday or Friday whenever it was
237 00:30:15.920 ⇒ 00:30:16.730 bencohen: like.
238 00:30:16.890 ⇒ 00:30:22.170 bencohen: We have a very, a very specific problem. Now, which is, we’re not selling enough of our top product.
239 00:30:22.620 ⇒ 00:30:24.889 bencohen: the variable speed line. So
240 00:30:26.090 ⇒ 00:30:38.380 bencohen: I have some suspicions as to what’s going on. But you guys will have an easier time waiting through the the shit to help figuring it out. We need to be very
241 00:30:38.540 ⇒ 00:30:51.280 bencohen: as you know, what’s going on in the macro sense of things we don’t have time for like a tam analysis, we can do it in parallel. But we need you guys to look directly into campaigns, pricing
242 00:30:51.380 ⇒ 00:30:58.210 bencohen: other things. You know, conversion, optimism, anything that could affect sales very specifically.
243 00:30:58.470 ⇒ 00:31:07.259 bencohen: This is all very good, and I have no problem working with you on it. But we we need to do tactical stuff right now. There’s no time
244 00:31:07.700 ⇒ 00:31:13.919 bencohen: to to spend a month of of investigation of the pool industry. We already know a lot of this stuff.
245 00:31:15.870 ⇒ 00:31:16.350 bencohen: So when.
246 00:31:16.350 ⇒ 00:31:17.169 Robert Tseng: Yeah, no, I mean.
247 00:31:17.170 ⇒ 00:31:19.489 bencohen: Very, very micro actually,
248 00:31:20.070 ⇒ 00:31:40.759 Robert Tseng: Great. Yeah, I mean, as much as you can kind of like help get a sense of priority of like, where do we want to drill into. Specifically, I think, for me, this is just the the framework that I apply to when I’m looking at what the opportunity for the business. So if you have a particular skew that you’re trying to optimize, and all effort is going to that, then great, we’ll we’ll go straight to that. And that’s that’s that’s the question that we should be answering.
249 00:31:41.070 ⇒ 00:31:46.520 Daniel Schonfeld: Yeah, Robert. This is Dan. Nice to meet you. By the way, and thanks thanks for putting this together.
250 00:31:47.520 ⇒ 00:31:50.319 Daniel Schonfeld: yeah, like, like Ben, said we, we it’s
251 00:31:51.250 ⇒ 00:32:07.159 Daniel Schonfeld: the the market is obviously always shifting. We we do have a. The bigger issue is not so much that it’s not selling. It’s that we have very little product coming in, due to tariffs which which I believe will subside. And and we’re obviously working on alternative manufacturing methods.
252 00:32:07.806 ⇒ 00:32:24.569 Daniel Schonfeld: But I see this as an opportunity point, while yes, it’s a scary time, and it’s difficult, because we don’t have a ton of goods on the water or an inventory. We probably have enough for the next 4 to 6 months, but this is a great time to regroup, rethink things. So
253 00:32:24.570 ⇒ 00:32:38.439 Daniel Schonfeld: when the tariffs subside which they will, and when we have alternative manufacturing which we will, you know our price point could change a bit like the tariffs could go from 1 45 down to 50, but that’s still double the tariffs that we had previously.
254 00:32:38.500 ⇒ 00:32:41.009 Daniel Schonfeld: So it’s a great time for us to rethink
255 00:32:41.120 ⇒ 00:33:08.900 Daniel Schonfeld: and reevaluate the data to try to understand. You know, how are we gonna operate in this new world profitably? And so I do think it’s it’s a good exercise to do. And I think what Ben’s alluding to is like we need. We do need to to kind of focus in on on a little bit more of the micro tactics and from our products set what’s really working, what’s really what’s not. And we also have. You know, I spoke to Ben last week we had kind of a powwow and other business development opportunities.
256 00:33:09.160 ⇒ 00:33:19.360 Daniel Schonfeld: The one thing that that we’re not taking advantage of is we we import slash manufacture nearly a thousand different skews. And we we haven’t really tapped into
257 00:33:19.510 ⇒ 00:33:33.479 Daniel Schonfeld: all the different opportunities as far as upsell bundles that our competitors do. And so we have access to any product pretty much in the pool space other than probably gas heaters. At the moment
258 00:33:33.600 ⇒ 00:33:38.589 Daniel Schonfeld: any any kind of knickknack upsell chemicals floats.
259 00:33:38.720 ⇒ 00:33:53.020 Daniel Schonfeld: My father-in-law just brought in 3 full containers of floats even at the higher tariff. So we’re trying to figure out, how do we get a higher Aov, you know just kind of tacking on to what we’re selling now.
260 00:33:53.300 ⇒ 00:34:00.410 Daniel Schonfeld: So those kind of things are helpful, like, what are other companies and what are other products that are being sold in the market
261 00:34:00.770 ⇒ 00:34:09.949 Daniel Schonfeld: being being bundled with or upsold. With that we could take advantage of, and now increase our margins, even when the tariffs if they go to 50%,
262 00:34:10.150 ⇒ 00:34:14.489 Daniel Schonfeld: right? So how do we add another 25% onto our margin?
263 00:34:14.639 ⇒ 00:34:40.069 Daniel Schonfeld: So those are the things we’re trying to evaluate. You know, there’s really only 2 ways to gain margin. It’s you increase the revenue or decrease the costs. We’re trying to look at the burn it at both sides a bit right now and again. This is an opportunistic time, not a time to panic or or freak out. It’s just it’s a time to be hyper, aware of what our costs are? And and are we doing everything we can to increase our order values? And so I think if you focus there.
264 00:34:40.090 ⇒ 00:34:44.840 Daniel Schonfeld: That would be very, very helpful to us, and really kind of affect the bottom line.
265 00:34:46.900 ⇒ 00:34:53.159 Robert Tseng: Yeah, no, I totally hear you. I think the yeah. The tam exercise aside, I think really just narrowing on, you know.
266 00:34:53.320 ⇒ 00:34:58.230 Robert Tseng: for your existing skews. Yeah, we could do some additional optimization to see, like
267 00:34:58.610 ⇒ 00:35:28.269 Robert Tseng: you know, what you can do to get to squeeze more out of that. But then I think your biggest lever is company is your ability to launch products. And how do you bundle things? I think that’s just consistent across that industry? And so knowing, like what skews you should be bundling together like, how, what trends do you see coming in your industry like what are like? How do you ride the right wave as you’re as you’re launching new products together. I think that’s really kind of like that. That would be where I would be spending the time to show you like. What are those different levers that you could, that you could be pulling.
268 00:35:28.590 ⇒ 00:35:30.159 bencohen: That’d be great. Yeah, yeah.
269 00:35:30.160 ⇒ 00:35:40.040 Daniel Schonfeld: I I agree with you there. We we haven’t done that in quite some time, nor have we really launched any new products in quite some time other than some new black and decker filters and other products.
270 00:35:40.610 ⇒ 00:35:56.750 Daniel Schonfeld: our ability at this moment to create brand new products is is virtually nonexistent until we till we get some relief on the tariff. So obviously, you know, equipment’s extremely expensive. Molds have to be made in order to change anything. But we do have access, like I said. And so
271 00:35:56.750 ⇒ 00:36:01.459 Daniel Schonfeld: exactly, yeah, I’m happy to provide even our existing inventory skews for
272 00:36:01.630 ⇒ 00:36:09.979 Daniel Schonfeld: for that we have in stock. Now we’ve got nearly 8 million dollars worth of non black and decker skews in our warehouses today.
273 00:36:10.280 ⇒ 00:36:13.530 Daniel Schonfeld: That’s sitting there waiting to be sold. And so
274 00:36:14.011 ⇒ 00:36:33.019 Daniel Schonfeld: we have retail stores also. But they’re obviously gonna gonna leverage that as well. But we are certainly welcome to tap into any of those skews. And so I’m happy to provide what those skews are. And you know they’re they’re. It’s quite robust. There’s probably a thousand skews that we have
275 00:36:33.310 ⇒ 00:36:46.089 Daniel Schonfeld: many, many of them non useful for this situation. But but there’s going to be quite a few gems in there, and if we can figure out, you know how to take. You know, when we sell a heat pump, it typically comes on a pallet.
276 00:36:46.230 ⇒ 00:36:54.800 Daniel Schonfeld: Could we stick 10 more items on that pallet and up the order value without adding anything to the shipping cost? The answer is definitely, yes, on that
277 00:36:54.950 ⇒ 00:36:59.699 Daniel Schonfeld: I don’t. But we we’ve got to test these things and figure out which which products are going to make sense
278 00:37:01.190 ⇒ 00:37:07.249 Daniel Schonfeld: and also so when we ship a pump, it may cost 30 to 40 bucks to ship it.
279 00:37:07.430 ⇒ 00:37:13.539 Daniel Schonfeld: and then, when we ship a bag of shock separately, a chemical bag of 24, it costs 20 bucks.
280 00:37:13.770 ⇒ 00:37:21.510 Daniel Schonfeld: but could we get if we bundled them together. Could we still sell it for the the both those items for the same amount?
281 00:37:22.050 ⇒ 00:37:38.280 Daniel Schonfeld: The pool pump would still cost 40 bucks, but maybe if they’re bundled together and attached together. The overall shipping cost is still lower than shipping, both of those individually, which is again just an exercise I’m trying to figure out. Those are answers. We can easily get through. Chuck at our warehouse.
282 00:37:38.784 ⇒ 00:38:06.650 Daniel Schonfeld: But that’s another win. And so, finding those those small wins whereby somebody was going to order those separately at a different point. Maybe they buy it together. We lower our shipping costs even by 5 bucks. It’s still additional margin. So I’m looking to find those kind of diamonds in the rough, those those average order value increases. So when times are good again, we’ve done all this work, you know now. And we’ve analyzed, we’ve we’ve kind of sharpened our pencils to figure out.
283 00:38:07.100 ⇒ 00:38:28.400 Daniel Schonfeld: you know, a better process for identifying bundles, new products. And so back to your original thing. I don’t want to dismiss a tam exercise. I do think it’s very important, and I think it’s important to understand the trends of the industry, I think what Ben was trying to say was, you know we are trying to get we? We don’t want to spend weeks doing it, which I don’t think is what you insinuated.
284 00:38:28.820 ⇒ 00:38:29.140 Robert Tseng: Yeah.
285 00:38:29.140 ⇒ 00:38:35.939 Daniel Schonfeld: And I do think it’s important. But we do need to start getting into more executable tasks, so we can start generating revenue now.
286 00:38:36.650 ⇒ 00:38:42.649 Uttam Kumaran: Yeah. So I think a good takeaway on our end is, yeah, you’re totally right. The theme of today is just like time box everything to like.
287 00:38:42.880 ⇒ 00:38:48.651 Uttam Kumaran: basically like a week or 2, where we could qualify whether we’ve it’s worth the squeeze. And then or.
288 00:38:48.940 ⇒ 00:38:53.879 bencohen: Maybe they can. I’m very fine with little Little, these little, these many little sprints.
289 00:38:54.330 ⇒ 00:38:54.670 Uttam Kumaran: Yeah.
290 00:38:54.670 ⇒ 00:39:03.120 bencohen: Figure out if we’re on the right track. If we are or not, we go to the next thing. But there’s another one I want to put on your radar like Tom when you, when you 1st came on
291 00:39:03.440 ⇒ 00:39:10.909 bencohen: year ago, or whatever it was, we made a ton of headway on shipping, and what happened is.
292 00:39:11.150 ⇒ 00:39:25.801 bencohen: at the same time we were onboarding more and more 3 pls to the point where all of that work that you did with getting ups. That crazy reduction it was. It didn’t even. It was moot. Because we were basically doing everything through 3 pls.
293 00:39:26.690 ⇒ 00:39:29.560 bencohen: we need to look at that again, because I actually think that
294 00:39:29.730 ⇒ 00:39:32.600 bencohen: we will. The 3 Pl. Stock
295 00:39:32.840 ⇒ 00:39:39.360 bencohen: is getting lower and our warehouse stock is getting higher. So we actually might be in a position again where we can
296 00:39:39.540 ⇒ 00:39:42.079 bencohen: push back harder there. But this is something to look at, because
297 00:39:42.620 ⇒ 00:39:48.247 bencohen: the 3 pls are they? They come with a different set of challenges with regard to pricing consistency.
298 00:39:48.560 ⇒ 00:39:49.100 Uttam Kumaran: You know.
299 00:39:49.400 ⇒ 00:39:50.510 bencohen: So.
300 00:39:51.010 ⇒ 00:39:54.390 Uttam Kumaran: Are you guys with the same ups person, Kelly? So Kelly.
301 00:39:54.480 ⇒ 00:39:56.539 bencohen: Yeah, yeah, yeah. Still, Cali, but
302 00:39:56.830 ⇒ 00:39:58.899 bencohen: we don’t have volume for them, because we.
303 00:39:59.260 ⇒ 00:40:01.870 Uttam Kumaran: So heavy we wanted. We wanted to de-risk.
304 00:40:01.870 ⇒ 00:40:06.620 Uttam Kumaran: We didn’t get the volume discounts that like at towards the back. Yeah.
305 00:40:06.620 ⇒ 00:40:14.529 bencohen: Yes, but you know, and that was a strategic company decision. We did not want to be overly reliant on our Long Island location.
306 00:40:15.159 ⇒ 00:40:21.870 bencohen: It cost us. But I mean, you know, there’s a there’s a there’s, there’s a positive cost, negative cost. But we need to look at this one again.
307 00:40:22.110 ⇒ 00:40:31.030 Uttam Kumaran: I mean the main lever on the other side was shipping speed right? So shipping speed, and we lowered the average zone to like 3 or 4 versus it was.
308 00:40:31.340 ⇒ 00:40:40.449 Uttam Kumaran: you know, a lot higher. It was like closer to 5 or 7. So, and for that objective, your goal is like, okay, can we go for another reduction from the
309 00:40:40.730 ⇒ 00:40:42.910 Uttam Kumaran: Yapank location?
310 00:40:42.910 ⇒ 00:40:46.230 Daniel Schonfeld: The heat pumps are a separate thing. We we can talk about that.
311 00:40:46.230 ⇒ 00:40:54.839 Daniel Schonfeld: Just let me let me chime in. Be careful with that, because the the number one driver. Yes, it was to gain more margin.
312 00:40:55.070 ⇒ 00:40:58.120 Daniel Schonfeld: But we want to be out of island wrecks.
313 00:40:58.120 ⇒ 00:40:58.750 Uttam Kumaran: Okay.
314 00:40:58.750 ⇒ 00:41:02.579 Daniel Schonfeld: Hole there like it’s a black hole.
315 00:41:03.220 ⇒ 00:41:23.529 Daniel Schonfeld: and it’s also free. It’s not real. And eventually it’s going to come to an end. And so when that does come to an end one day we don’t want to be scratching our heads, saying, Shit! We got to move everything out of there. Find 3 pls. The goal was to test it. But yeah, right now, it’s definitely makes sense to kind of sharpen our pencils. See if we can ship maybe a few more goods out of there.
316 00:41:23.840 ⇒ 00:41:35.819 Daniel Schonfeld: because we’re we’re in a tight spot for the next few months, and we really need to keep costs low. But we still want to keep those balls in the air, because ultimately the goal is to become independent of island Rec and their warehousing.
317 00:41:38.740 ⇒ 00:42:03.339 Daniel Schonfeld: what else is, I gonna say, oh, the the also, the we have to remember the reduction of warranty issues, I think, came into play because we were shipping heaters directly to Florida. It was a few less trips and and few less. you know, in and out of the containers that the the heat, the heat pumps had to take, which reduced some of the the issues with punctures and and things breaking. So
318 00:42:03.360 ⇒ 00:42:11.860 Daniel Schonfeld: we just have to be cognizant of all those things, while while reducing costs, is very, very important and probably paramount. We do have to keep those other things into account, because they all add up.
319 00:42:12.340 ⇒ 00:42:16.809 Uttam Kumaran: So let’s just I mean, we’ll do something on just showing what the data looks like for shipping costs.
320 00:42:16.810 ⇒ 00:42:17.520 bencohen: Yeah.
321 00:42:17.680 ⇒ 00:42:21.649 Uttam Kumaran: Back from when we did the reduction to post 3 Pl.
322 00:42:22.070 ⇒ 00:42:26.250 Uttam Kumaran: And you’ll see it on like a unit level and like an aggregate level.
323 00:42:26.820 ⇒ 00:42:27.310 Daniel Schonfeld: Yup!
324 00:42:27.460 ⇒ 00:42:34.879 bencohen: Yeah, the like. There’s certain you gotta you gotta break it by class. So like, if you look at like the pump class
325 00:42:35.210 ⇒ 00:42:37.823 bencohen: like the water pump, not the heater pump.
326 00:42:38.590 ⇒ 00:42:48.119 bencohen: Those costs with the 3 pl. Have been relatively successful where we lost control and big time was on heaters. The the 1st problem was
327 00:42:48.860 ⇒ 00:42:55.400 bencohen: a bunch of the carriers that Eunice engaged early on for heat pumps went out of business.
328 00:42:55.790 ⇒ 00:42:58.870 bencohen: This is in the Biden time. So this is, you know.
329 00:42:59.370 ⇒ 00:43:05.729 bencohen: the drama was a year ago, or something like that. So basically, when we had, like a deal where anywhere in the State of Florida. It would be a hundred
330 00:43:06.110 ⇒ 00:43:09.319 bencohen: $5 to land it to somebody’s thing.
331 00:43:09.810 ⇒ 00:43:15.239 bencohen: it just it just evaporated. And they they went. They opted
332 00:43:15.370 ⇒ 00:43:22.470 bencohen: for kind of safer, bigger options, which is fine. It’s more reliable because the customers is always right, but
333 00:43:22.870 ⇒ 00:43:25.910 bencohen: it wiped away margin so.
334 00:43:26.110 ⇒ 00:43:31.290 bencohen: And there’s like a few other cascading problems similar to that. So we just got to look at it all.
335 00:43:32.890 ⇒ 00:43:33.650 bencohen: yeah.
336 00:43:34.180 ⇒ 00:43:53.700 bencohen: we can do. Not. By the way, Robert Tom, I’m I’m around. We can. You probably have another. You probably already missed your 10 o’clock calls my guess because you scheduled 30 so if you want, just tell me when you need me today, I give you an hour just we can get into. We can go, you know, little deeper into all these things. So we have a real plan to attack.
337 00:43:54.390 ⇒ 00:43:55.340 Robert Tseng: Okay. Yeah.
338 00:43:55.340 ⇒ 00:43:55.730 Daniel Schonfeld: Hello!
339 00:43:55.730 ⇒ 00:43:57.550 Robert Tseng: I don’t go ahead.
340 00:43:57.990 ⇒ 00:44:08.660 Daniel Schonfeld: I was just gonna say, what’s your degree of confidence with the the skewed data? That project that we did. I would like to start using that data.
341 00:44:08.660 ⇒ 00:44:12.190 Uttam Kumaran: Yeah, I mean, we. I feel like we made a lot of headway. I was sort of
342 00:44:12.380 ⇒ 00:44:16.500 Uttam Kumaran: waiting a lot more on for your lead on pushing that forward like
343 00:44:16.830 ⇒ 00:44:28.570 Uttam Kumaran: we got that said like in a great place. There’s like there’s like there’s still like a couple of outstanding, like smaller things to figure out, but we were sort of we. It’s sort of on pause right now.
344 00:44:28.970 ⇒ 00:44:34.830 Daniel Schonfeld: Okay, I’m gonna go. I’ll take a look at it in the next hour. Just see where we are. I know there was a Google Sheet, with all the different tabs and.
345 00:44:34.830 ⇒ 00:44:38.620 Uttam Kumaran: Yeah, take a look at the Google Sheet. If you want to bump up the last email we sent.
346 00:44:38.900 ⇒ 00:44:44.629 Daniel Schonfeld: That would be good. That would be good. And then, are you guys, Ben, are you actively using the, the
347 00:44:44.750 ⇒ 00:44:45.910 Daniel Schonfeld: the real system.
348 00:44:45.910 ⇒ 00:44:54.870 bencohen: I am now that it’s been. They spent so much time the last 3 weeks updating it. But before that it wasn’t accurate enough to be used. But now I am. Yeah.
349 00:44:54.870 ⇒ 00:44:59.299 Uttam Kumaran: Yeah, we’re actively. We’re actively talking to Kim almost every week.
350 00:44:59.600 ⇒ 00:44:59.900 bencohen: Yeah.
351 00:45:00.200 ⇒ 00:45:01.710 Uttam Kumaran: Most days, so.
352 00:45:02.120 ⇒ 00:45:02.550 Daniel Schonfeld: Okay.
353 00:45:02.770 ⇒ 00:45:03.500 Uttam Kumaran: Too.
354 00:45:03.500 ⇒ 00:45:08.719 Daniel Schonfeld: Can, it would, along with that email, can you resend me the link my login.
355 00:45:08.990 ⇒ 00:45:09.710 Uttam Kumaran: Sure.
356 00:45:09.710 ⇒ 00:45:11.550 Daniel Schonfeld: I’m going to take a look through it also today.
357 00:45:11.730 ⇒ 00:45:12.420 Uttam Kumaran: Sure.
358 00:45:12.830 ⇒ 00:45:18.197 Daniel Schonfeld: That’s that’s all. I got guys. And and Robert really, really nice to meet you. I think Amber’s on.
359 00:45:18.860 ⇒ 00:45:22.910 Daniel Schonfeld: Thank you guys for getting on is amber. A is amber. A person.
360 00:45:23.190 ⇒ 00:45:24.430 Uttam Kumaran: Yes.
361 00:45:26.005 ⇒ 00:45:26.360 Amber Lin: Indeed!
362 00:45:26.360 ⇒ 00:45:27.633 Daniel Schonfeld: Okay in.
363 00:45:28.110 ⇒ 00:45:28.649 Daniel Schonfeld: I don’t know.
364 00:45:28.650 ⇒ 00:45:31.099 Uttam Kumaran: I was like a it was like a perfect. It’s a perfect.
365 00:45:31.465 ⇒ 00:45:34.754 Uttam Kumaran: So we’re not. We’re not that advanced yet, but.
366 00:45:35.120 ⇒ 00:45:37.910 Daniel Schonfeld: I know you guys are pushing the the limits here of technology.
367 00:45:37.910 ⇒ 00:45:44.879 Uttam Kumaran: Hopefully, we’ll be out of, I mean, hopefully, we’ll have been out of the business by that. Otherwise, like I don’t know what the business is. There’s no people, I guess. I don’t know.
368 00:45:45.400 ⇒ 00:45:47.312 Uttam Kumaran: Everything’s an agent. Yes,
369 00:45:47.790 ⇒ 00:45:49.370 Daniel Schonfeld: Sorry amber
370 00:45:51.270 ⇒ 00:46:07.599 Daniel Schonfeld: anyways. Thank you, guys, I do appreciate you. And I I appreciate you sticking by us with in this difficult time. Obviously, the next bunch of months are gonna be very challenging with everything going on on the macro issues. But we’ll we’ll get through it. The company’s been around 50 years. We’ve got a lot of inventory to to weather the storm and
371 00:46:07.886 ⇒ 00:46:25.590 Daniel Schonfeld: you know, the the thing I told you about Utam is not going away. The private equity deal. They’ve they’re up my butt actually still engaged still paying for legal fees. So there! That that is going to happen at some point. And you know, at the at the end of the day we’re in no rush. We we didn’t need to sell
372 00:46:25.720 ⇒ 00:46:46.200 Daniel Schonfeld: or get this investment this year. It’s more for a boost. It’s exciting opportunity. But it’s not going anywhere. And neither are we. So I appreciate you guys. And yeah, I’m excited. I’m really excited for some of these projects to come to fruition. I think the app could be really be really be exciting. We’re not relying on it, but I think it’s got a lot of opportunity.
373 00:46:46.200 ⇒ 00:46:52.219 Uttam Kumaran: No, I think you’re probably the only people in pool that can do it like actually get around to like doing something.
374 00:46:52.360 ⇒ 00:47:07.179 Daniel Schonfeld: Yeah, it would be really neat, and it’s got a lot of a lot of exciting prospects if we can get it right. I understand. It’s a very long term project to get to where we’d love it for it to be. It’s years and a lot of money. But we got to start somewhere, and now is a really good time.
375 00:47:07.180 ⇒ 00:47:07.800 bencohen: Yeah.
376 00:47:08.320 ⇒ 00:47:08.870 Uttam Kumaran: Okay.
377 00:47:09.250 ⇒ 00:47:10.690 Daniel Schonfeld: All right. Thank you. Guys.
378 00:47:11.230 ⇒ 00:47:12.170 Uttam Kumaran: Thanks. Everyone.
379 00:47:12.640 ⇒ 00:47:13.769 Robert Tseng: Talk things around.