Brainforge Knowledge

LMNT_Account_Transition_Prep

10 min read

LMNT Account Transition Preparation - Robert as Account Manager

Created: February 16, 2026
Purpose: Comprehensive recall document combining Awaish discussion (Feb 13) with recent Shivani meetings (Feb 10-13) to prepare for transition to account manager role

🎯 Goal

Get Robert caught up and prepared to drive productive working sessions with Shivani, reducing Uttam’s involvement in random calls and ad-hoc requests.

🔑 Access & Tools Needed

Snowflake Access

  • Status: Need own Snowflake account/credentials (not using Uttam’s)
  • Action: Awaish to send invite to create new user with credits
  • Issue: Robert experiencing table loading errors when trying to access data
  • Note: Awaish mentioned this is a Snowflake-side error, not access-related

Data Access Overview

  • Polytomic: Ingests data to Snowflake
  • dbt: Runs transformations in Snowflake
  • Raw Data Location: RAW.POLYTOPIC.SHOPIFY schema
    • Orders, order line items, customers from Shopify
    • Wholesale orders tagged with “wholesale” string in tags field
  • Retail Data: External share from Amazon (12 tables for Walmart, 3 for Target)
  • Documentation: Available in Cursor/BrainForge vault

Transcripts & Documentation

  • Status: All transcripts should be automatically going to LMNT info in vault
  • Note: Awaish requested this feature, Uttam said he’s working on it
  • Action: Robert needs time to review all transcripts to reference previous discussions

đź“Š Current State of LMNT Project

Data Engineering Architecture

Data Sources

  1. Shopify (via Polytomic)

    • E-commerce orders
    • Wholesale orders (tagged with “wholesale”)
    • Both come from same Shopify source

  2. Retail Data (External shares)

    • Walmart: 12 tables (POS data + order management system data)
    • Target: 3 tables
    • Amazon: Shared data (separate from Polytomic)

  3. CRM Data

    • Wholesale team CRM lives in Google Sheets
    • Combined with Shopify data in models

Data Models (BroadMarts Schema)

Wholesale:

  • wholesale_mart - Weekly views for wholesale
  • wholesale_dm_customer - Combines Shopify + CRM data
  • All wholesale partners/accounts

Retail:

  • retail_mart - Retail reporting
  • Two key tables:
    • fact_sales - POS data from both Walmart and Target
    • retail_fact_walmart_only_sales - Walmart order management system data (NOT POS)
  • Note: For total retail revenue, need to join both tables together (Awaish working on this)

E-commerce:

  • Not yet modeled in product mods
  • Only wholesale and retail are in BroadMarts currently

👤 Understanding Shivani’s Working Style & Concerns

Key Behavioral Patterns

  1. Over-Reliance on BrainForge

    • Doesn’t have vision/future state she’s working towards
    • Very over-reliant on BrainForge team
    • Struggles to put together basic things (does basic stuff in Google Sheets)
    • Uttam’s assessment: “She’s out of her depth but cares a lot about winning and optics”

  2. Micro-Management Tendencies

    • Wants to be in the middle - doesn’t want BrainForge going directly to stakeholders
    • Approves meetings with wholesale team before they happen
    • Frequently changes priorities:
      • Said to pause wholesale development, move to retail
      • Then suddenly asks for new wholesale requests
    • Wants to QA everything live with the team

  3. Communication Style

    • From Feb 11 meeting: Wants purposeful meetings with specific topics
    • Gets frustrated if meetings don’t have concrete things to discuss
    • Prefers working sessions where she can pull up data and compare sources
    • Asks many questions about data differences and reconciliation

  4. Quality vs Speed Tension

    • Sets high bar for quality/formatting
    • Then asks for quick turnaround (“sprint mode”)
    • Uttam’s feedback: “I can’t get anything in front of you until it’s perfectly formatted because that’s what you asked for”
    • Shivani’s response: “We should be able to hash things out”
    • Robert’s take: Need to set guardrails on iteration process

Key Concerns & Questions She Raises

  1. Data Reconciliation

    • Always asking: “How is what we’re showing different from what they think it is?”
    • Compares BrainForge data to:
      • Finance/accounting systems (we don’t have access)
      • Previous work experience/structures
      • Internal Element resources
    • Issue: Revenue doesn’t match finance (<10% difference)
    • Context: Wholesale team has no issues with our reporting; finance is the problem
    • Reconciliation work: Amber working on this based on sheets Shivani sent to Uttam

  2. Data Source Differences

    • Questions about Walmart vs Target data:
      • Are they apples-to-apples?
      • How do they define “week ending” dates?
      • What are the differences in how they recognize revenue?
      • What discrepancies exist when joining sources?
    • Robert’s insight: She wants us to proactively call out these differences

  3. Usability Questions

    • “Can I actually use what you give me?”
    • “What can I glean from this retail data right now?”
    • Focuses on whether people at LMNT can answer basic questions
    • Gap: Mismatch between progress updates (Gantt charts) and her need to actually use the data

🚧 Current Challenges & Blockers

Process Issues

  1. Handoff Problems

    • Work is siloed: ETL → Modeling → Reporting
    • Robert can’t easily jump in and reference what Awaish is doing
    • Need better internal sandbox/environment for collaboration

  2. Iteration Process

    • No clear agreement on how to do iterations
    • Shivani wants to hash things out live
    • Team needs guardrails on when to do quick iterations vs formal reviews

  3. Stakeholder Access

    • Shivani controls access to stakeholders (wholesale team, Will from retail)
    • Pattern: One discovery call → build → show → get feedback → iterate
    • But Shivani wants to approve all direct contact

Technical Issues

  1. Revenue Definitions

    • Need to align on definitions (e.g., “net revenue” = revenue minus discounts/refunds?)
    • Haven’t had follow-up conversations with Will (retail) after initial discovery
    • Similar pattern happened with wholesale - built first, then got feedback on calculations

  2. Data Joining Challenges

    • Walmart POS + Walmart order management system need to be joined
    • Different data sources may recognize revenue differently
    • Need to document and explain differences

đź“… Upcoming Work & Priorities

Immediate (Next Week)

  1. Ingestion Work

    • Meta, TikTok, Snapchat data (planning next week)
    • Amazon and Walmart connectors (blocked by Polytomic building them)
    • Salesforce data
    • Shivani’s goal: Ingest as much as possible in parallel

  2. BI Tool (Omni)

    • Demo happening this week
    • Trial instance starting
    • Anticipated ask: Duplicate all Sheet reporting into Omni
    • Timeline: Next week will be “pressing” with ingestion + Omni starting simultaneously

  3. Snowflake AI Analyst

    • Uttam wants to give Shivani access
    • Awaish sent demo clip to Uttam
    • Will be enabled on Prod Marts with docs loaded
    • Purpose: Keep her occupied before Omni is ready

Ongoing Work Streams

Three lanes:

  1. Wholesale - Weekly views, customer models
  2. Retail - POS data, revenue reporting
  3. Biz Ops - Broader reporting at LMNT (how they do reporting)

đź’ˇ Strategies for Productive Sessions

Setting Expectations

  1. Challenge Finance Mismatches

    • “Accounting recognizes revenue differently than we do - it’s not going to match”
    • This is common across clients
    • <10% difference is not a problem

  2. Establish Iteration Process

    • Can’t give final-state quality every time
    • Need agreement on when quick iterations are appropriate vs formal reviews
    • Reference Eden model: “They have to let us do our work”

  3. Focus on Usability

    • Help her understand how to use what we’re building
    • Bridge gap between progress updates and actual usability
    • Once Omni is ready, she can play with it directly

Meeting Structure

From Feb 11 meeting learnings:

  • Shivani wants purposeful meetings with specific topics
  • Come prepared with:
    • What we’ve learned from retail so far
    • What we’re seeing in the data
    • Differences between sources (Walmart vs Target)
    • Ability to answer: “What were sales across both stores yesterday?”
  • Don’t schedule meetings just to check status - need concrete things to discuss

Key Questions to Be Ready For

  1. Data Source Questions:

    • How are Walmart and Target data different?
    • What discrepancies exist when joining sources?
    • How do they define dates/revenue differently?

  2. Usability Questions:

    • Can I answer basic questions with this data?
    • What can I actually glean from this right now?
    • How do I use this to answer stakeholder questions?

  3. Reconciliation Questions:

    • Why doesn’t this match finance?
    • How are you doing reconciliations?
    • What’s the difference between our data and their accounting systems?

🎯 Action Items for Robert

Before Monday Meeting

  1. âś… Get Snowflake access (Awaish sending invite)
  2. âś… Review all transcripts in vault (spend weekend time)
  3. âś… Meet with Awaish to walk through Snowflake structure
  4. ✅ Understand current models and what’s available
  5. ✅ Prepare to answer Shivani’s common questions

During Monday Meeting (2 hours)

  1. Set Expectations:

    • Establish iteration process
    • Set guardrails on when quick vs formal reviews
    • Challenge finance mismatch expectations

  2. Understand Her Needs:

    • What questions is she trying to answer?
    • What’s blocking her from using the data?
    • What does she need to be successful?

  3. Demonstrate Capability:

    • Show ability to pull up data and discuss it
    • Answer questions about data differences
    • Help her understand how to use what we’ve built

Ongoing

  1. Reduce Uttam’s Load:

    • Take over ad-hoc calls
    • Drive working sessions
    • Be the point person for reporting requests

  2. Improve Internal Process:

    • Better collaboration with Awaish
    • Create sandbox environment for data access
    • Improve handoff between stages

  3. Help Shivani:

    • Bridge gap between progress and usability
    • Teach her how to use the data
    • Set up Snowflake AI Analyst as interim solution

đź“ť Key Quotes & Insights

From Uttam (Feb 12 meeting):

  • “She occupies a lot of our time… I can go make more money for us elsewhere”
  • “There’s no other consulting partner that’s gonna do this type of work with this type of person”
  • “She’s really out of her depth but cares a lot about winning and optics”
  • “Either we become the punching bag, or she just has this insecurity”
  • Pushing her to hire someone from Brave to be “the punching bag”

From Shivani (Feb 11 meeting):

  • “If we’re gonna jam, then we need something to jam about”
  • “What is gonna make all of this work successful is if it lands for the people at LMNT to be able to answer basic questions”
  • “I’m not trying to be difficult, I’m just… if we’re gonna jam and do these sessions, then I need a topic at hand”

From Awaish (Feb 13 meeting):

  • “She wants us to ingest as much as possible”
  • “Next week is going to be pressing… we will start with ingestion, then Omni, and everything at the same time”
  • “She is in the middle… she don’t want us to go directly with anyone”

🔄 Workflow Pattern with Shivani

Typical Cycle:

  1. Discovery call with stakeholder (wholesale team, Will from retail)
  2. Build initial models/tables based on what we have
  3. Show to stakeholder → get feedback
  4. Iterate based on feedback (calculations, fields, structure)
  5. Shivani reviews → compares to finance/other sources
  6. Reconciliation work if needed
  7. Move to next domain → repeat

Current State:

  • Wholesale: Built, shown, iterated, now paused (new requests go to backlog)
  • Retail: In progress - building models, will show to Will soon
  • E-commerce: Not yet started in product mods

🚨 Red Flags & Watch Outs

  1. Scope Creep: Shivani says to pause wholesale, then asks for new wholesale requests
  2. Quality vs Speed: Sets high bar, then wants quick turnaround
  3. Micro-management: Wants to control all stakeholder access
  4. Finance Mismatch: Will likely continue to be an issue - need to set expectations
  5. Over-reliance: She doesn’t know how to use the data herself - need to teach her

âś… Success Criteria

  1. Robert can:

    • Pull up any table/data in Snowflake during calls
    • Answer questions about data differences and sources
    • Help Shivani understand how to use the data
    • Drive productive working sessions without Uttam

  2. Shivani:

    • Can answer basic questions with the data
    • Understands how to use what we’ve built
    • Stops pulling Uttam into random calls
    • Has clear process for iterations and requests

  3. Team:

    • Better collaboration between Robert and Awaish
    • Clear handoff process
    • Reduced time spent on ad-hoc requests
    • More efficient working sessions

đź“š Reference Documents

  • Comprehensive Analysis: See LMNT_Uttam_Shivani_Thought_Partner_Analysis.md for detailed analysis of how Uttam guides Shivani and how to transition that role
  • All transcripts in: /knowledge/clients/unassigned/transcripts/
  • Recent meetings:
    • 2026-02-13: LMNT Data Engineering Sync (Awaish walkthrough)
    • 2026-02-12: LMNT Project Updates and Planning
    • 2026-02-11: Brainforge x Shivani Reporting Sync
    • 2026-02-11: LMNT Reporting and Tracking Sync
    • 2026-02-10: Eden and LMNT Reporting Discussion
  • Key historical meetings:
    • 2025-12-30: LMNT January Planning (extensive strategic planning session)
    • 2025-12-09: Element Client Engagement Planning Sync
    • 2025-11-18: Brainforge x LMNT Next Steps
    • 2025-10-23: Uttam <> Shivani (initial discovery call)

This document should be updated after Monday’s meeting with Shivani to capture new insights and adjust strategies.

Graph View