Top N SKU Model

FreshServe transforms access to food in urban food deserts. We deliver fresh meal kits through a community-driven model and a strategic focus on a data-driven Top N SKU portfolio to scale impact.

Comment

Inspiration

FreshServe’s mission is clear: fight urban food deserts with affordable, fresh meal kits that actually improve diets. Recent signals made the opportunity urgent and actionable: strong community outcomes despite cost pressure, and internal feedback that operations needed a frontline-first reset. We set out to convert that mission traction into a durable operating moat by simplifying assortment to the smallest high-impact set of SKUs, improving recipe usability, and shifting toward asset-light logistics to reduce leverage risk and expand margins.

What it does

Top N SKU Model is a repeatable, data-driven process that selects the smallest set of products that delivers the majority of demand while respecting kitchen capacity, pantry constraints, and community nutrition needs.

  • Scores each SKU by economic throughput and operational simplicity, then ranks them.
  • Produces a cumulative coverage curve that shows exactly how many SKUs are needed to reach a target share of orders, typically 75 to 85 percent.
  • Enforces a pantry cap so the selected portfolio does not overload storage or create spoilage risk.
  • Outputs a Recommended N and a feasibility flag for S&OP to act on each month.
  • Integrates with the broader FreshServe plan: menu engineering, pictogram-first recipes, and asset-light route selection using 3PL on low-density corridors.

How we built it

We implemented the model as an auditable Excel workbook, so operations leaders can own the process without custom software.

  1. Data inputs

    • Last 13 weeks of units sold per SKU, standard cost elements, and six complexity drivers: unique ingredients, steps, allergen handling, setup time, forecast error, and supplier reliability.

  2. Scoring model

    • Contribution margin per SKU: Price − (Food + Packaging + Variable Labor + Spoilage).
    • Throughput: Contribution Margin × Units.
    • Complexity composite: weighted blend of normalized drivers.
    • Final Score: Throughput × (1 − Alpha × ComplexityComposite).
    • Ranks SKUs by Score and computes cumulative coverage and cumulative ingredients.

  3. Decision rule and outputs

    • Coverage_N: smallest N with Coverage ≥ target.
    • IngredientCap_N: largest N that fits the pantry cap.
    • Recommended_N: MAX(Coverage_N, IngredientCap_N) plus a feasibility message.
    • Coverage_Curve sheet renders an investor-ready chart.
    • Implemented in the workbook FreshServe_SKU_Selection_Template.xlsx for immediate use.

  4. Operating fit

    • Ties into FreshServe’s current flow: local sourcing, urban kitchen assembly, QC on freshness and portions, refrigerated distribution, and SMS feedback. The model strengthens each step by reducing variety where it does not add value and by concentrating throughput on the winners.

Challenges we ran into

  • Ingredient de-duplication
    The simple pantry cap may overcount unique ingredients across SKUs. For the MVP we used a conservative sum. A follow-on enhancement will compute a true set union for exact pantry size.
  • Recipe usability and portion clarity
    Customers reported confusion. We coupled Top N with pictogram-first cards, timers, and portion color-coding to reduce support contacts and churn.
  • Fleet fixed costs and leverage
    Owned refrigerated vehicles increased gearing to 70 percent. The plan introduces 3PL partners on low-density routes and reserves owned fleet for densest corridors to lower capital intensity.
  • Culture and change management
    Anonymous feedback highlighted morale and burnout. We embedded Gemba, Kaizen, 5S, takt time, and a right-to-stop-the-line policy so teams lead improvements rather than absorb top-down pressure.

Accomplishments that we're proud of

  • A clear, defensible selection method that leadership can explain to crews, partners, and investors. It replaces opinion with a transparent score and a visible curve.
  • A ready-to-use tool that any site lead can run monthly without engineers. It prints the Recommended N and flags feasibility in plain language.
  • Direct line to outcomes
    The model supports FreshServe’s social mission by protecting affordability and improving reliability in food deserts while building financial resilience. Community data already show 90 percent reporting better diets and 85 percent reporting affordability.
  • Capital-discipline story
    Top N underpins the board narrative: fewer SKUs with higher density, lower waste, and a path to deleveraging and margin expansion.

What we learned

  • Less can be more
    Concentrating volume into a compact, high-overlap pantry simplifies procurement, stabilizes yield, and raises first-pass quality.
  • Usability is an economic lever
    Clear instructions and portion signals reduce waste and increase repeat orders.
  • Frontline-first wins trust
    When operators help set N, define takt time, and surface hazards, morale improves and the system sustains itself. Internal voice must shape the plan, not just marketing.
  • Asset-light increases options
    Partnering on low-density routes lowers fixed costs and risk without giving up control where density is strongest.

What's next for Top N SKU Model

  1. Exact pantry union
    Add an ingredient set-union calculator so the pantry cap reflects true overlap. I can deliver a Python-assisted generator or an Excel macro version on request.
  2. Quarterly S&OP integration
    Make Top N a standing agenda item with automated refresh and a Power BI dashboard for coverage, pantry size, and per-drop costs.
  3. 3PL performance telemetry
    Roll out temperature and on-time SLAs with scorecards to ensure the asset-light mix protects quality and cost.
  4. Community co-design and A/B tests
    Iterate pictogram cards and portion tiers with community panels and SMS micro-surveys. Track Net Easy Score weekly using the existing channel.
  5. Investor-ready exhibits
    Include the Coverage Curve, Recommended N, and route density maps in each board and investor update for transparency and confidence.

Future development and digital transformation

What if FreshServe Ltd launched a dedicated Copilot, built with Microsoft Copilot Studio, that operations leaders, kitchen teams, and partner managers could talk to like a teammate? In a single prompt, managers ask: “What Top N SKU set will reach 80 percent coverage this week within our pantry cap, and what should we buy and ship?” The Copilot scores SKUs, generates the coverage curve, recommends N, drafts supplier commitments, proposes 3PL route mixes for low‑density corridors, and outputs pictogram recipe updates and SMS nudges. Leadership sees target outcomes from the strategic plan become day‑to‑day workflows: lower COGS, less spoilage, clearer instructions, improved retention, and a path to reduce gearing from 70 percent toward 45 to 50 percent while scaling impact in food deserts.

Built With

Share this project:

Updates