Meeting Title: Brainforge x Eden Strategy Overview Date: 2025-08-11 Meeting participants: Robert Tseng
WEBVTT
1 00:00:27.150 ⇒ 00:00:34.420 Robert Tseng: Okay, recording this video just to talk through everything that we do with Eden. …
2 00:00:37.150 ⇒ 00:00:40.790 Robert Tseng: So maybe a good place to start would be to talk through
3 00:00:42.640 ⇒ 00:00:51.200 Robert Tseng: kind of the strategy that we put together for them. So, this is already kind of more of a Q2 to Q3 strategy, but…
4 00:00:51.490 ⇒ 00:01:04.050 Robert Tseng: Really the main objectives are, one, to consolidate fragmented data tools. With any CPG company, yeah, at the end of the day, you’re basically running a
5 00:01:04.970 ⇒ 00:01:11.190 Robert Tseng: you’re running a marketing company. Most of your growth is driven through
6 00:01:11.430 ⇒ 00:01:16.640 Robert Tseng: Paid ads and, digital channels.
7 00:01:16.860 ⇒ 00:01:20.680 Robert Tseng: Trying to get your products distributed, you know.
8 00:01:21.040 ⇒ 00:01:25.419 Robert Tseng: Across the major platforms, if you’re able to.
9 00:01:26.260 ⇒ 00:01:33.980 Robert Tseng: … And I think, on average, something like 20-plus, you know, tools, that marketers are looking at.
10 00:01:35.430 ⇒ 00:01:46.010 Robert Tseng: Getting everything into a single place to have one comprehensive view of marketing and reporting is not common, or it’s not an easy task to achieve.
11 00:01:46.350 ⇒ 00:01:58.190 Robert Tseng: And so what ends up happening is that there’s a lot of manual stitching, where, marketing teams basically log into a bunch of different tools, and they have to copy-paste
12 00:01:58.300 ⇒ 00:02:07.559 Robert Tseng: into, you know, Google Sheets or spreadsheets, used by the team through these CSV exports.
13 00:02:09.039 ⇒ 00:02:16.660 Robert Tseng: This data is pretty static, it’s stale in that, it’s only relevant to the time of your export.
14 00:02:17.180 ⇒ 00:02:25.699 Robert Tseng: And you can’t really get a sense of freshness and accuracy when you’re spending all your time just creating a bunch of different spreadsheets.
15 00:02:26.230 ⇒ 00:02:33.820 Robert Tseng: And so… … maybe something I can share is…
16 00:02:37.550 ⇒ 00:02:39.390 Robert Tseng: I’ll spend…
17 00:02:46.950 ⇒ 00:02:48.099 Robert Tseng: Aye, aye, aye.
18 00:03:11.590 ⇒ 00:03:16.069 Robert Tseng: Yeah, you get situations like this, where this was actually broken. I gotta share it.
19 00:03:34.360 ⇒ 00:03:35.350 Robert Tseng: …
20 00:03:36.440 ⇒ 00:03:44.510 Robert Tseng: Yeah, you can see that this was really just a spend sheet put together for a bunch of different data sources. Even as I’m pulling it up, if one
21 00:03:44.800 ⇒ 00:03:53.330 Robert Tseng: little thing breaks in the formatting, the entire spreadsheet breaks, and so right now, we’re kind of in the dark with regards to this data.
22 00:03:58.410 ⇒ 00:04:03.500 Robert Tseng: And… So, in order to drive further consolidation.
23 00:04:03.620 ⇒ 00:04:08.380 Robert Tseng: We had to put together, like, an actual data architecture diagram.
24 00:04:11.640 ⇒ 00:04:16.650 Robert Tseng: and build out a more robust data platform. So, we can kind of zoom into this.
25 00:04:16.940 ⇒ 00:04:28.680 Robert Tseng: And so, here are just a couple examples where you have raw data sources. Northbeam, Zendesk, BASC, Google Sheets. BASC is a telehealth platform, similar to Shopify.
26 00:04:29.000 ⇒ 00:04:29.880 Robert Tseng: …
27 00:04:30.540 ⇒ 00:04:39.230 Robert Tseng: There’s a lot more marketing data sources that aren’t visualized here, but the point is just to show the differences across the different data ingestion tools.
28 00:04:39.380 ⇒ 00:04:54.050 Robert Tseng: And so, we have some data sources that are coming in just through custom Google Cloud Functions, which was what we had set up before. Then we have other data sources that are coming in through Polytomic and through Segment.
29 00:04:54.200 ⇒ 00:04:59.179 Robert Tseng: Segment isn’t mod isn’t, isn’t, demonstrated here, but, …
30 00:05:00.230 ⇒ 00:05:08.160 Robert Tseng: These are more out-of-the-box, ETL data connector tools that are able to move data from one place to another.
31 00:05:08.360 ⇒ 00:05:19.539 Robert Tseng: They bring them into BigQuery, and from BigQuery, we’re able to build out a whole set of different models. So we have SalesMark, so these are all the different models that, a…
32 00:05:20.330 ⇒ 00:05:24.549 Robert Tseng: Sales reporting, would, would need.
33 00:05:24.740 ⇒ 00:05:34.609 Robert Tseng: You know, you would want a customer’s table, transactions, shipments, there are treatments as well, and different types of summarized views, for…
34 00:05:35.400 ⇒ 00:05:40.819 Robert Tseng: for Eden’s product portfolio, so that they can have all of the
35 00:05:41.000 ⇒ 00:05:50.219 Robert Tseng: downstream reports, which are built out in Tableau, kind of visible and usable, for the end user.
36 00:05:50.760 ⇒ 00:06:00.050 Robert Tseng: You have the marketing mart, so this is really for the, marketing team to be able to get a really good view of marketing
37 00:06:00.720 ⇒ 00:06:07.020 Robert Tseng: allocations, something that they care about a lot is called the Marketing Efficiency Ratio, or MER.
38 00:06:07.230 ⇒ 00:06:20.010 Robert Tseng: Where you consider all, online, offline channel spend and performance, and you try to get that all into one place so that you can understand, how efficiently you’re spending your marketing dollars.
39 00:06:20.960 ⇒ 00:06:29.210 Robert Tseng: Then there’s a customer support piece, where from an operational perspective, your orders are going through a lot of different journeys, and this needs to really be updated.
40 00:06:29.400 ⇒ 00:06:35.540 Robert Tseng: So I’m gonna… Take this, and we’re gonna actually… Here.
41 00:06:38.490 ⇒ 00:06:39.270 Robert Tseng: Okay.
42 00:07:21.120 ⇒ 00:07:24.080 Robert Tseng: Okay, … Yeah.
43 00:07:27.200 ⇒ 00:07:31.040 Robert Tseng: So, coming back here, that helps us to…
44 00:07:31.250 ⇒ 00:07:40.639 Robert Tseng: Build out the architecture that will consolidate fragmented data tools so that all data now is centralized and living in a single place in the data warehouse.
45 00:07:40.770 ⇒ 00:07:42.800 Robert Tseng: And there’s a clear…
46 00:07:45.180 ⇒ 00:07:58.020 Robert Tseng: There are clear checks and balances in place to, monitor the health of this pipeline, make sure that data is fresh, that data is being transformed into the different,
47 00:07:58.300 ⇒ 00:08:01.899 Robert Tseng: Models that, power reporting.
48 00:08:03.160 ⇒ 00:08:19.610 Robert Tseng: Second is that Eden is preparing to transition to their own EMR system. And so, they recently acquired, a software company to reduce their dependency on Basque Health.
49 00:08:20.470 ⇒ 00:08:23.840 Robert Tseng: … this EMR system is…
50 00:08:24.010 ⇒ 00:08:35.910 Robert Tseng: gonna be their order management system, their employee medical record system. It’s gonna capture everything about patients and orders and transactions all in a single place. And this system is a backend architecture
51 00:08:36.120 ⇒ 00:08:51.619 Robert Tseng: that is going to, to connect to the data warehouse and all of the other tools that rely on it for sales reporting, transactions reporting. Right now, Basque has only been publishing webhooks. They’ve done that intentionally because they want
52 00:08:51.750 ⇒ 00:08:54.569 Robert Tseng: Their customers to be operating within their platform.
53 00:08:54.600 ⇒ 00:09:14.080 Robert Tseng: and yet the platform is kind of rigid and doesn’t produce all the reporting needs that they need. And so, there’s been just this back and forth on frustration that you do need a custom system built for your business, rather than just trying to continue to scale off of something that’s a product built out of the box that doesn’t have all the customizations that you would want.
54 00:09:15.310 ⇒ 00:09:16.280 Robert Tseng: …
55 00:09:16.810 ⇒ 00:09:25.109 Robert Tseng: And then thirdly, the main… the third objective is full funnel visibility into patient journey. So, prior to Brainforge working with this company.
56 00:09:25.260 ⇒ 00:09:31.460 Robert Tseng: They didn’t really know what was happening, in the holistic customer…
57 00:09:31.840 ⇒ 00:09:39.379 Robert Tseng: what they call a patient journey. But since working with Brainforge, we now have a single customer data model,
58 00:09:43.050 ⇒ 00:09:44.870 Robert Tseng: Which I can show here.
59 00:09:59.110 ⇒ 00:10:00.000 Robert Tseng: Oops
60 00:10:03.270 ⇒ 00:10:04.790 Robert Tseng: I’m calling…
61 00:10:16.410 ⇒ 00:10:17.560 Robert Tseng: …
62 00:10:26.400 ⇒ 00:10:28.590 Robert Tseng: So let me just roll the screen real quick.
63 00:10:36.320 ⇒ 00:10:40.649 Robert Tseng: Oh yeah, we have every customer, all the details, when they signed up.
64 00:10:41.150 ⇒ 00:10:43.420 Robert Tseng: where they came in from, UTMs.
65 00:10:44.980 ⇒ 00:10:49.659 Robert Tseng: kind of… whether they clicked via an advertisement on Google or Facebook.
66 00:10:50.230 ⇒ 00:10:55.490 Robert Tseng: What questionnaire they ended up taking in order for us to collect some of their, personal information.
67 00:10:55.730 ⇒ 00:10:56.760 Robert Tseng: …
68 00:10:59.970 ⇒ 00:11:04.970 Robert Tseng: How many orders they spent, they’ve placed, how much money they’ve spent, their first visit.
69 00:11:05.080 ⇒ 00:11:08.820 Robert Tseng: The transactions… …
70 00:11:13.330 ⇒ 00:11:22.950 Robert Tseng: And the specific order, the product that they ordered, how many products they’ve ordered, just truly an end-to-end view with all the different custom milestones.
71 00:11:24.020 ⇒ 00:11:25.539 Robert Tseng: Captured in a single mop.
72 00:11:28.990 ⇒ 00:11:30.090 Robert Tseng: …
73 00:11:35.030 ⇒ 00:11:35.810 Robert Tseng: Yeah.
74 00:11:36.120 ⇒ 00:11:42.100 Robert Tseng: And… So let’s move into some of the reports that we’ve built out.
75 00:11:44.740 ⇒ 00:11:49.300 Robert Tseng: So one such dashboard is a executive-level marketing dashboard.
76 00:11:49.580 ⇒ 00:11:50.400 Robert Tseng: Where…
77 00:11:50.580 ⇒ 00:12:06.840 Robert Tseng: We have it broken out by product and order, and we’re able to, do cohort LTV, LTV cap ratio, so we can see the lifetime value of a customer since their first order, kind of grouped by the different
78 00:12:06.970 ⇒ 00:12:14.980 Robert Tseng: purchase… order purchase months, from the first time they purchase. We have to look at trends, growth and revenue trends.
79 00:12:15.390 ⇒ 00:12:17.470 Robert Tseng: So that we can understand.
80 00:12:17.680 ⇒ 00:12:29.130 Robert Tseng: When does… what… you know, understand some sort of seasonality over the past 6 months? Where have we seen the most revenue growth? Where has revenue dipped?
81 00:12:29.490 ⇒ 00:12:30.400 Robert Tseng: And…
82 00:12:31.680 ⇒ 00:12:43.040 Robert Tseng: Yeah, just that full marketing efficiency ratio, so that we can know how efficient, marketing spend is across paid and unpaid channels, and including all the different online and offline spends.
83 00:12:45.260 ⇒ 00:12:48.319 Robert Tseng: Other dashboards that we have, …
84 00:12:54.270 ⇒ 00:12:57.079 Robert Tseng: This one is super important for the ops team.
85 00:12:57.520 ⇒ 00:13:05.529 Robert Tseng: So, this gives them an ability to look at processing time order statuses, and to have the highest level order
86 00:13:05.640 ⇒ 00:13:12.240 Robert Tseng: operational KPIs. So, order to ship date, so when a customer places an order to when it actually got shipped.
87 00:13:12.370 ⇒ 00:13:23.099 Robert Tseng: order to pharmacy, and when an order was placed, to when it was sent to the pharmacy or the fulfillment provider, or pharmacy to ship to order delivered. So we have some SLAs for, like, when, how…
88 00:13:23.340 ⇒ 00:13:29.999 Robert Tseng: what the turnaround time should be for, between these different stages. And, we can… we’ve noticed that
89 00:13:30.290 ⇒ 00:13:31.290 Robert Tseng: …
90 00:13:31.330 ⇒ 00:13:44.470 Robert Tseng: This turnaround time has, is being impacted on a daily, weekly basis, based on some of the operational adjustments. And so, the ops team really relies on this report to be able to figure out, which orders are at risk.
91 00:13:44.470 ⇒ 00:13:52.210 Robert Tseng: which orders, have been shipped out of SLA that need to be brought up to pharmacies to hold them accountable to
92 00:13:52.220 ⇒ 00:13:55.129 Robert Tseng: For why they didn’t actually ship them out on time.
93 00:13:56.250 ⇒ 00:13:59.590 Robert Tseng: Out of SLA orders, so when orders
94 00:13:59.590 ⇒ 00:14:16.999 Robert Tseng: are super delayed, extremely delayed, and maybe they’re… they’re stuck for some reason, that’s another escalation point. So just being able to help the ops team really figure out what are all the different escalation points where they need to be given some sort of signal and direction for when they… where they need to go and
95 00:14:17.000 ⇒ 00:14:20.770 Robert Tseng: Take action, with… with fulfillment providers.
96 00:14:20.770 ⇒ 00:14:28.380 Robert Tseng: This is an incredibly, incredibly helpful tool, help, for the ops team. We’ll manage performance of just…
97 00:14:28.770 ⇒ 00:14:37.249 Robert Tseng: you know, orders. When you’re transacting with a customer, the main thing you care about is making sure that you get the order that you wanted to order on time.
98 00:14:37.300 ⇒ 00:14:51.459 Robert Tseng: And if there are any delays, that’s going to be the bulk of the customer service issues. People are always going to be asking about where the order is. And so, having just this ability to, break down the customer… the order journey step-by-step.
99 00:14:51.460 ⇒ 00:15:01.000 Robert Tseng: And to be able to look at all of these different inflection points, is extremely helpful for the team to, understand what’s happening to their orders.
100 00:15:02.910 ⇒ 00:15:05.939 Robert Tseng: Beyond that, we have…
101 00:15:06.350 ⇒ 00:15:17.619 Robert Tseng: executive-level reporting as well, and so one example would be product ROAS and LTV snapshots. We’ll go there, and we’ll also queue up the…
102 00:15:17.900 ⇒ 00:15:19.420 Robert Tseng: Dashboard.
103 00:15:23.780 ⇒ 00:15:32.230 Robert Tseng: Yeah, so this is just at a high level. Product level, revenue and, marketing performance, and so…
104 00:15:32.500 ⇒ 00:15:51.529 Robert Tseng: baked into all these different marketing platforms, you’re only looking at performance at a channel or campaign level, because that’s how you set up, the ads. You’re not really, categorizing them by product, and so that requires stitching of marketing platform data with internal,
105 00:15:53.970 ⇒ 00:15:59.909 Robert Tseng: Just product-level data as well, making sure that, you know, we’re able to group
106 00:16:00.080 ⇒ 00:16:03.220 Robert Tseng: Campaigns and channels,
107 00:16:04.040 ⇒ 00:16:21.450 Robert Tseng: to create different levels of aggregation, and that needs to be done in the custom way in the data warehouse. And so, any way that executives want to look at the business, this is extremely important to look at product-level profitability, which products are, you know, contributing the most revenue to the overall portfolio.
108 00:16:21.620 ⇒ 00:16:22.530 Robert Tseng: …
109 00:16:23.130 ⇒ 00:16:35.760 Robert Tseng: This is extremely important for when you’re launching new products and wanting to know the efficacy of a particular product. And so, this is not something that you could, look at just by relying purely on the data that you’re seeing within marketing platforms.
110 00:16:37.070 ⇒ 00:16:47.729 Robert Tseng: We also have LTV CAC breakdown, so being able to calculate specific financial metrics so that we understand where, customers are able to
111 00:16:50.430 ⇒ 00:16:54.930 Robert Tseng: At what point do customers pay back, the costs that it…
112 00:16:55.150 ⇒ 00:17:03.610 Robert Tseng: to acquire them. That’s important for marketers, so that they know how much money they should be spending, what they should be expecting, spending in ROI.
113 00:17:03.840 ⇒ 00:17:05.470 Robert Tseng: On their investment.
114 00:17:05.720 ⇒ 00:17:18.190 Robert Tseng: The more that we can predict these, revenue metrics at a patient level or customer level, we’re able to run more targeted ads, it helps with setting pricing expectations.
115 00:17:18.349 ⇒ 00:17:23.080 Robert Tseng: And we can ensure that the bundles and offers that we position
116 00:17:23.640 ⇒ 00:17:30.889 Robert Tseng: Help reduce the amount of time that it takes customers to pay back, the cost of their investment.
117 00:17:31.040 ⇒ 00:17:35.789 Robert Tseng: And we know how much they need to pay back in order for them to be a high-value customer for us.
118 00:17:38.070 ⇒ 00:17:44.620 Robert Tseng: So there’s all kinds of just revenue and sales, related,
119 00:17:45.620 ⇒ 00:17:52.520 Robert Tseng: Reporting, that’s, helpful for constructing, The best offers possible.
120 00:18:04.870 ⇒ 00:18:09.000 Robert Tseng: Also, take a look at… …
121 00:18:17.020 ⇒ 00:18:18.570 Robert Tseng: attention dashboard.
122 00:18:19.940 ⇒ 00:18:32.689 Robert Tseng: Yeah, not just from an LTV perspective, but… for a subscription, product, … How many customers are…
123 00:18:32.900 ⇒ 00:18:52.620 Robert Tseng: sticking with us after their first purchase, and for how long? We should see a steady drop-off rate over a period of time, but we should be able to compare across monthly cohorts. Are we better at retaining customers, or getting worse at retaining customers? And so that’s an extremely important metric for this subscription business to… to measure.
124 00:18:53.890 ⇒ 00:18:54.910 Robert Tseng: …
125 00:18:58.970 ⇒ 00:18:59.810 Robert Tseng: Yeah.
126 00:19:03.050 ⇒ 00:19:11.129 Robert Tseng: Okay, so I think, you know, just kind of going back to strategically, you know, we’re able to,
127 00:19:11.970 ⇒ 00:19:16.670 Robert Tseng: Make Customer Journey Insights more self-service, because now that we have
128 00:19:16.750 ⇒ 00:19:33.590 Robert Tseng: a fully enriched customer data model that lives in the data warehouse, stitches all the possible customer touchpoints, and everything we know about a customer into a single model. We’re able to push that into behavioral analytics tools and lifecycle platforms, like Mixpanel and Customer IO.
129 00:19:33.860 ⇒ 00:19:35.940 Robert Tseng: Where those…
130 00:19:36.090 ⇒ 00:19:48.530 Robert Tseng: the data that we push into those tools is able to enrich the level of segmentation that we’re able to create. It’s no longer just off of basic things, such as order place, when was the last order placed, number of orders.
131 00:19:48.530 ⇒ 00:19:57.570 Robert Tseng: But we’re able to look at some behavioral and demographic things at a more granular level. The products they purchase, time since last purchase.
132 00:19:57.690 ⇒ 00:20:10.649 Robert Tseng: whether or not they… how many tickets that they’ve submitted through customer support requests. And so we can be more creative about constructing these segments of what a high-value customer looks like, and then being able to load them up into
133 00:20:10.650 ⇒ 00:20:18.380 Robert Tseng: a customer I.O. or Braze platform, so that we can have extremely targeted messaging to help win them back if they’ve gone dormant.
134 00:20:18.380 ⇒ 00:20:34.139 Robert Tseng: or if they’re a high-value customer, we know they’re going to spend more, because we’ve set… we’ve seen that their predicted LTV is higher, by… and they’re, you know, less than halfway to fulfilling that potential. We have opportunities to continue to sell them new products.
135 00:20:34.140 ⇒ 00:20:38.440 Robert Tseng: To bump them up to their… to help them maximize the revenue potential.
136 00:20:39.120 ⇒ 00:20:57.460 Robert Tseng: And so, being able to unlock cohort-based analysis for marketers is a game changer. They’re able to be creative and go in there and learn about all the different cuts of their customers so that they can better optimize the campaigns that they run against them.
137 00:21:01.730 ⇒ 00:21:16.279 Robert Tseng: And then, yeah, just having a data infrastructure that’s agnostic to whatever system helps Eden to be able to migrate EMR providers without worrying about it too much.
138 00:21:16.710 ⇒ 00:21:19.630 Robert Tseng: What we have to do is kind of re…
139 00:21:20.550 ⇒ 00:21:33.789 Robert Tseng: instrument the telemetry for mapping schemas between one system versus another, but from an analytics perspective, none of the downstream reporting is going to break, because all of that still lives on top of the same models. And so.
140 00:21:33.790 ⇒ 00:21:47.500 Robert Tseng: Yeah, the data modeling that you do once is… it’s not just set and forget. There is some maintenance required, but it’s not something that you have to continue to upgrade. It is world-class. It is how the world’s best companies build their data… run their data functions already.
141 00:21:47.880 ⇒ 00:21:48.660 Robert Tseng: …
142 00:21:49.300 ⇒ 00:22:05.139 Robert Tseng: So overall, I think the point is not to just create a bunch of infrastructure. There’s a lot of consolidation that happens. We want to speed up time to insight. That’s the main metric that we’re going for. We want to be able to operate within the constraints of being a lean data team.
143 00:22:05.140 ⇒ 00:22:12.319 Robert Tseng: For the price of, like, one to two senior data hires, you’re going to be able to run a complete data function
144 00:22:12.320 ⇒ 00:22:21.280 Robert Tseng: That, prioritizes against the needs of the business, and knows what the most important,
145 00:22:21.320 ⇒ 00:22:31.109 Robert Tseng: KPIs to move are for a high-revenue business that’s going from 100 million plus. The more that we can
146 00:22:31.110 ⇒ 00:22:37.090 Robert Tseng: focus on getting all the basics, basic reporting set up and automated, then analysts
147 00:22:37.090 ⇒ 00:22:52.329 Robert Tseng: can… then… then the time shifts to doing deeper dive analysis, predictive modeling, machine learning, and AI-driven analysis work, that’s going to be able to empower operators and decision makers to, make better decisions within the workflows that they’re already using.
148 00:22:52.780 ⇒ 00:22:54.280 Robert Tseng: Okay, that’s it.