2025-12-05_carlos_ecommerce

Meeting Agenda: Carlos (E-commerce) Discovery Call

Date: Thursday, December 4, 2025 (Meeting occurred Dec 4, agenda dated Dec 5) Time: 10:00 AM PT (1.5 hours)
Attendees: Carlos (LMNT E-commerce Manager), Uttam Kumaran, Awaish Kumar, Shivani Amar

---

POST-MEETING SUMMARY

Key Learnings

Current State:

• Carlos manually compiles E-commerce Growth Dashboard from 10+ data sources - spends estimated 10-15 hours/month on reporting
• Dashboard structure: Blue cells = manual inputs, Black cells = calculated outputs (this color coding system works well for Carlos)
• Currently forecasting is “target-setting” not true forecasting - backing into annual revenue targets rather than bottoms-up projections
• Team recently switched to this new reporting format (November 2025) which revealed data quality issues not visible in old template

Data Sources & Systems:

• Shopify: Uses Source Medium for AOV and new vs. returning customers, GA4 for traffic (sessions), manually categorizes revenue into 12+ channels (Subscriptions, Email, Direct, Organic, SEO, paid channels)
• Amazon: Pulls Buy Box % from Business Reports, traffic from Sales & Traffic Report, new customers from Source Medium (hashed email matching - has 15-30 day latency issue)
• Walmart: Minimal revenue (“non-material”), pulls AOV from Analytics Console, infers new customers from “Repeat Order %” inverse (workaround - no direct metric available)
• Ad Platforms: Direct exports from Meta Ads Manager, Google Ads Console, Amazon DSP, Tatari (linear TV agency)
• Source Medium: 90-95% confidence level, uses GA4 attribution (last-click), but Carlos has found calculation errors requiring manual QA

Key Pain Points Discovered:

• Amazon New Customer Latency: When switching to new template in November, discovered 30-50% variance in new customer counts due to hashed email matching lag. Now waits 4+ days after month-end before finalizing numbers
• GA4 Limitations: Cannot separate drink mix vs. sparkling traffic for channel attribution (needed for Meta analysis)
• Walmart Data Quality: No direct new customer metric, must use last 30 days repeat % as inverse proxy (inaccurate)
• Forecasting Input Gaps: Wants to forecast using inputs (traffic, conversion rate, AOV, CPM) but can’t get channel-level traffic + conversion rate breakdown, forced to use revenue as input
• Source Medium Trust: Reports have errors occasionally, team prefers spreadsheets. Discrepancy tracking during GA4 transition caused 3-month reliability gap

Quick Wins Identified:

• Automate daily/monthly data pulls via API ingestion (eliminate 10-15 hrs/month manual work)
• Fix Amazon new customer reporting with snapshot + revised reporting pattern (known issue with solution path)
• Build proper bottoms-up forecasting model using traffic, conversion rate, and AOV as inputs (not just target-setting)
• Create unified metric definitions to eliminate confusion (3 different CAC definitions, multiple ROAS/MER variations)
• Address GA4 drink mix vs. sparkling separation issue for better channel attribution

Answers to Key Questions

Q1-4: Current Reporting & Processes

• Q1: Weekly routine

  • A: Daily tracking in coordinator-managed spreadsheet, monthly compilation for company-wide OKR tracker, pulls from 10+ sources

• Q2: Monthly contribution time

  • A: Estimated 10-15 hours compiling E-commerce section of James’s monthly report (significant burden)

• Q4: Manual data work

  • A: Pulling metrics from Amazon Seller Central (Buy Box %, traffic, orders), Walmart Seller Center, Source Medium, GA4, all ad platforms

Q5-10: Shopify Deep Dive

• Q5: Daily metrics

  • A: Total revenue, orders, conversion rate, AOV, traffic (sessions), new vs. returning customer split

• Q6: Customer funnel tracking

  • A: Uses GA4 for sessions, Source Medium for attribution, conversion rate calculated as orders/sessions

• Q7: Conversion rate investigation

  • A: October dropped 34.7% MoM (4.9% → 3.2%), using MER (Marketing Efficiency Ratio) to validate performance despite attribution gaps, still investigating root cause

• Q8: Subscription tracking

  • A: Source Medium tracks subscriptions as separate channel, uses “order sequence” filter to identify first vs. repeat purchases

• Q9: Refunds/returns

  • A: Not detailed in current conversation, needs follow-up

Q11-15: Amazon Deep Dive

• Q11: Regular Amazon reports

  • A: Business Reports (Buy Box %), Sales & Traffic Report (traffic), Source Medium (new customers via hashed emails)

• Q12: Advertising tracking

  • A: Tracks attributed revenue split by new vs. returning customers to calculate CAC separately from total ROAS

• Q13: Buy Box monitoring

  • A: Manual monthly pull from Business Reports “Featured Offer %” field, critical metric (1% Buy Box loss ≈ 1% revenue loss)

• Q14: Subscribe & Save

  • A: Tracked as separate non-paid channel alongside “Other Amazon Sales”

• Q15: Amazon data challenges

  • A: NEW CUSTOMER LATENCY IS MAJOR ISSUE - takes 15-30 days for proper tagging, causes huge variance at month-end (30-50% spike discovered in November)

Q16-18: Walmart.com

• Q16: Walmart vs. Amazon differences

  • A: Much more limited data availability, revenue is “non-material” so lower priority

• Q17: Available reports

  • A: Analytics Console (AOV), Customer Insights (repeat order % - use inverse as new customer proxy), Item Sales Report (traffic)

• Q18: Unique challenges

  • A: No direct new customer metric (biggest gap), can’t filter by exact dates (must use last 7/30 days windows), may include reseller traffic in page views

Q19-22: Marketing Attribution

• Q19: Channel tracking

  • A: Source Medium uses GA4 last-click attribution, UTM structure critical (breaks attribution if changed)

• Q20-21: Partnerships attribution

  • A: Blake’s partnership data flows through Source Medium using source/medium format (e.g., “andrewhuberman / sponsor”), tracked separately as channel

• Q22: Incrementality measurement

  • A: Started working with MMM (Marketing Mix Modeling) vendor because seeing cross-channel effects (Meta ad → Amazon purchase) that last-click attribution misses

Q22-27: Metrics & Definitions

• Q22: Revenue vs. sales

  • A: Used interchangeably, Source Medium uses “revenue,” Carlos’s dashboard uses “sales”

• Q23: CAC variations

  • A: Blended CAC = total performance spend / new customers across all channels (Carlos’s preferred method to avoid attribution issues)

• Q24: ROAS & MER

  • A: MER (Marketing Efficiency Ratio) = total E-com revenue / all spend, used as “true guiding light” because attribution isn’t perfect. Performance targets (ROAS) based on platform attribution

• Q25: Advertising-to-Sales Ratio

  • A: Essentially inverse of MER

• Q26: New vs. returning

  • A: Source Medium uses “order sequence” filter to identify first-time vs. repeat customers

Q26-29: Data Quality & Trust

• Q26: Most trusted sources

  • A: Platform native data (Shopify, Amazon, Walmart) for revenue/orders, 90-95% confidence in Source Medium but requires manual QA

• Q27: Source Medium discrepancies

  • A: Yes, found calculation errors in reports requiring manual validation, GA4 transition caused 3-month unreliable period

• Q28: Validation checks

  • A: Cross-references Source Medium against platform exports, identified issues by comparing historical trends when switched to new template

• Q29: Manual reconciliation

  • A: Monthly at minimum, sometimes more frequently when discrepancies found

Q30-33: Pain Points

• Q30: Top automation priority

  • A: “Now you’re talking” - eliminate all manual data pulls from 10+ consoles (Amazon, Walmart, Source Medium, GA4, ad platforms)

• Q31: Recent struggle

  • A: Amazon new customer lag discovery (November), conversion rate drop diagnosis (October)

• Q32: Time wasters

  • A: Daily coordinator pulling data, monthly compilation of dashboard from multiple sources

• Q33: Blind spots

  • A: Can’t separate drink mix vs. sparkling traffic in GA4 for Meta attribution, can’t get channel-level traffic + conversion rate for forecasting

Q34: Third-Party Vendors

• Q34: Vendor roles

  • A: Need to follow up on specifics, Tatari confirmed as linear TV buying agency

Q35-37: Future Vision

• Q35: Ideal dashboard

  • A: All data automated in single platform (considering moving from Source Medium to fully owned solution), wants input-based forecasting not target-setting

• Q36: Helpful analyses

  • A: True bottoms-up forecasts using traffic/conversion/AOV inputs, better cross-channel attribution (MMM underway), channel-specific traffic + conversion breakdown

• Q37: Time reallocation

  • A: Strategic analysis instead of data compilation, investigating anomalies like conversion rate drops

Decisions Made

• Carlos is comfortable with current dashboard structure (blue = inputs, black = calculations) and doesn’t need format changes
• Team agreed forecasting should shift from “target-setting” to true projections using proper inputs (traffic, conversion, AOV)
• Acknowledged MER as “true guiding light” metric for overall efficiency vs. attributed ROAS for channel-specific optimization
• Carlos open to replacing Source Medium if Brainforge can provide better solution, but wants to see consistency in data sources maintained

Follow-Up Needed

• Carlos: Share 2025 actual reporting template (more accurate than 2026 planning version after 1 month of real data) (Due: Dec 10)
• Carlos: Provide access to Source Medium dashboards
• Carlos: Share sample Shopify exports showing 12+ channel breakdowns
• Carlos: Clarify third-party vendor roles (WBX, JG Marketing specifics)
• Brainforge: Present ETL options that support all Carlos’s data sources (Shopify, Amazon, Walmart, ad platforms) (Due: Dec 15)
• Brainforge: Evaluate Source Medium coverage vs. cost and provide recommendation on keep/replace (Due: Dec 20)
• Brainforge: Design snapshot + revised reporting pattern for Amazon new customer latency issue (Due: Jan 5)
• Brainforge: Build forecasting model prototype using traffic/conversion/AOV inputs (Due: Jan 15)
• Brainforge: Research solutions for GA4 drink mix vs. sparkling traffic separation
• Awaish: Follow up with Carlos via Slack for system access and file sharing (Due: Dec 6)
• Shivani: Create Slack channel for Carlos + Brainforge team (or add to existing)

---

Meeting Objectives

1. Deep dive into e-commerce data sources (Shopify, Amazon, Walmart.com)
2. Understand current reporting processes and pain points
3. Review the e-commerce growth dashboard and data sources
4. Identify quick wins and automation opportunities
5. Document KPIs and metric definitions for e-commerce channels
6. Understand relationship with Source Medium and data trust issues

---

Demos/Walkthroughs Requested

• E-commerce Growth Dashboard walkthrough (2026 Planning Dashboard)

  • Walk through the 7 separate tabs structure we see in your dashboard
  • How you populate this each month (manual vs. automated)
  • Time spent maintaining this per week/month (100+ metrics tracked)
  • Monthly delta (Δ) calculations - manual or formula-based?
  • Which sections you trust most/least

• Shopify reporting process

  • Walk through the 12+ sub-segments (Subscriptions, Email, Direct, Organic, Partnerships, Meta, TikTok, Google, Bing, etc.)
  • What reports do you pull from Shopify Admin?
  • What data comes from Source Medium (“SourceMedium New Customers”)?
  • Subscriber tracking: How do you track 400K+ subscriber target and monthly ship rate?
  • Data validation process

• Amazon reporting process

  • Walk through Amazon’s 4 sub-segments (Subscribe & Save, Other sales, Search, DSP)
  • Subscribe & Save tracking (281K Drink Mix + 48K Sparkling subscribers)
  • Buy Box Win Rate monitoring (94% target - how tracked?)
  • Amazon Advertising (Search vs. DSP split)
  • Monthly Ship Rate tracking

• Source Medium usage

  • Which Source Medium dashboards do you use?
  • Where does “SourceMedium New Customers” data come from?
  • What works well vs. what doesn’t?
  • Where do gaps exist?

---

Questions to Ask

Current Reporting & Processes

1. Walk me through your typical weekly reporting routine:

   • What reports do you create?
   • Who consumes them?
   • How long does it take?

2. For the monthly report James sends out, what’s your contribution process?

   • How long does it take to compile your section?
   • What’s the most time-consuming part?

3. What reports or analyses do you wish you could create but can’t today?

4. Where do you spend the most time on manual data work?

Shopify Deep Dive

5. What metrics do you track daily from Shopify?

   • Revenue/sales
   • Orders (new vs. subscription)
   • Conversion rate
   • AOV (average order value)
   • Traffic/sessions
   • Others?

6. How do you track the customer funnel?

   • Sessions → Product page views → Add to cart → Checkout → Purchase
   • Where is this data coming from?

7. How are you currently investigating the conversion rate drop (Oct 2025: 4.9% → 3.2%)?

   • What hypotheses have you explored?
   • What data would help answer this?

8. How do you track subscription vs. one-time purchases?

   • Active INSIDER subscribers
   • Churn and retention metrics
   • Subscription revenue attribution

9. How are refunds and returns handled in your reporting?

   • Net revenue calculation
   • Return rates by product/channel

10. What’s your process for tracking promotional performance?

    • Discount codes
    • Free sample programs
    • Give-A-Salt campaign tracking

Amazon Deep Dive

11. What Amazon reports do you use regularly?

    • Sales & orders
    • Subscribe & Save
    • Advertising (AMS/PPC)
    • Inventory levels

12. How do you track Amazon advertising performance?

    • ROAS (return on ad spend)
    • ACOS (advertising cost of sale)
    • Attribution window

13. Buy Box monitoring - how is this currently tracked?

    • Automated or manual?
    • Historical tracking?

14. How do you handle Amazon’s Subscribe & Save separately from one-time purchases?

15. What’s challenging about Amazon data?

    • Delays in reporting?
    • Reconciliation issues?
    • Missing attribution?

Walmart.com

16. How different is Walmart.com tracking from Amazon?
17. What reports are available?
18. Any unique challenges with Walmart data?

Marketing Attribution & Partnerships

19. How do you track which marketing channels drive Shopify sales?

    • UTM parameters?
    • Source Medium attribution?
    • First-touch vs. last-touch?

20. Partnerships Revenue Line Item:

    • I see “Partnerships” as a separate Shopify sub-segment with $25.7M budget (2026)
    • New Customers Ad Attributed Net Sales tracked separately
    • How does Blake’s partnership data flow into your dashboard?
    • Do you reconcile with Blake’s 100+ partner tracking spreadsheet?
    • Attribution: $5.7MtargetYTD,CAC$225

21. How is partnership revenue attributed? (e.g., Huberman, Dan Go)

    • I see Source/Medium format: “andrewhuberman / sponsor”
    • Is this UTM-based?
    • Promo codes also used?
    • How do you handle multi-touch (person sees podcast, then clicks Meta ad)?

22. How do you measure incrementality of paid social (Meta, TikTok)?

    • I see Meta budget: $6.4M for Shopify 2026
    • TikTok: $1.3M budget
    • Any A/B testing or holdout groups?

Metrics & Definitions

22. How do you define “revenue” vs. “sales”?

    • I see “Sales” as primary term in dashboard - same as revenue?
    • Gross vs. net of refunds?
    • When is revenue recognized (order date vs. ship date)?

23. CAC Definitions - I see three different ones:

    • “eCom Blended CAC (first time purchases)” - target $20.61
    • “Ad-attributed CAC”
    • “Carlos CAC (Carlos Only)”
    • What’s the difference? Which one is the “real” CAC?

24. ROAS & MER Definitions - I see multiple:

    • “Ad-attributed ROAS”
    • “LC ROAS” (Last-Click ROAS)
    • “MER” (Marketing Efficiency Ratio) - target 2.4
    • “Acquisition MER”
    • How do these relate to each other? Are they measuring the same thing?

25. “Advertising-to-Sales Ratio” - Is this just inverse of MER?

26. What’s your definition of “new customer” vs. “returning customer”?

    • How is “Returning Customer Sales” attributed to paid media?
    • ”% Share of Total Paid Media Orders” - how calculated?

27. What KPIs are you measured on?

    • I see $436M e-commerce target for 2026
    • Monthly targets and YoY growth
    • Subscriber count targets (400K Shopify, 281K Amazon Drink Mix)

Data Quality & Trust

26. Which data sources do you trust most? Least? Why?

27. Have you found discrepancies between Source Medium and raw data?

28. What validation checks do you run on your reports?

29. How often do you have to manually reconcile numbers?

Pain Points & Priorities

30. If you could automate one thing, what would it be?

31. What question did you recently struggle to answer?

32. What takes up your time that feels like it shouldn’t?

33. Where do you feel “blind” in your reporting?

Third-Party Support & Agencies

34. I see three third-party vendors in your dashboard:
    • WBX Third Party Support ($563K annual)
    • JG Marketing Third Party Support ($330K)
    • Tatari Third Party Support ($396K for TV/streaming)
    • What do each of these vendors do for you?
    • How do they integrate with your reporting?

Future Vision

35. What does your ideal e-commerce dashboard look like?

36. What analyses would help you make better decisions?

37. How would you like to spend your time if manual reporting was automated?

---

Resources Mentioned/Requested

To be filled out during/after meeting

• Access to live e-commerce growth dashboard (2026 Planning version)
• Source Medium login credentials
• Sample Shopify exports (showing all 12+ sub-segments)
• Sample Amazon Seller Central reports (Subscribe & Save, Buy Box)
• Documentation on CAC calculation methodologies (all 3 versions)
• Documentation on ROAS/MER calculations
• Sample of monthly delta calculation process

---

Action Items

To be filled out during/after meeting

Carlos

• [ ]

Brainforge (Uttam/Awaish)

• [ ]

Shivani

• [ ]

---

Notes

Space for additional notes during the meeting

Key Insights

Red Flags / Concerns

Quick Win Opportunities