Postobón Profitability Comparison
Proposal for Postobón · Negotiation tool · Traditional trade channel

The numbers come from the shopkeeper.

Today the shopkeeper decides what to display and what to buy based on whichever product they believe earns them the most. This tool turns that conversation into a numeric exercise: four data points per product —dictated or typed— and the real profit of the competitor versus its Postobón counterpart, in pesos and as a percentage, right in their own store.

Try the working demoSee the tour
Scroll to walk through the visit
The visit walkthrough

Eight steps, one single question: which one earns more?

Tap any step to see what happens, what the rep says, and what gets logged for BI.

Tap a step to see the detail
1
Step 1 of 8

    How it works

    From the conversation in the store
    to the purchase decision.

    01

    Captures what the shopkeeper says

    Purchase price, discount, sale price and boxes per month, for the competitor's product and its Postobón counterpart. The rep dictates it by voice or types it in; price lists come preloaded and are editable in the field.

    02

    Calculates the real profit

    Not gross sales: the actual profit. Net cost after discount, unit margin and monthly profit for each side, over the same horizon. The difference in pesos and as a percentage appears on a single screen, readable from a meter away.

    03

    Logs the session with no extra effort

    Point of sale, duration, SKUs compared, data, result and closing survey are all saved automatically. That data feeds the BI dashboard and the Business analytics.

    SKU comparison

    Units per month
    boxes × units_per_box
    Net cost
    purchase_price × (1 − discount/100)
    Unit margin
    sale_price − net_cost
    Monthly profit
    units × unit_margin
    Difference
    Postobón_profit − competitor_profit

    Combo comparison

    Food cost
    quantity × unit_price
    Beverage cost
    purchase_price × (1 − discount/100)
    Profit per combo
    combo_price − (food + beverage)
    Monthly profit
    profit_per_combo × combos_per_month
    Postobón scenario
    only the beverage is swapped; same price, same food item
    Boxes of 24 units except Gatorade and Agua Cristal (12): units per box is an attribute of the SKU, not of the system
    Module 5 · AI voice capture

    The rep talks.
    The fields fill themselves in.

    The rep repeats what the shopkeeper tells them. The engine transcribes the audio, a language model extracts each data point, assigns it to the right product and puts it in its field. The rep just confirms. Here's what it looks like:

    Lucro VoiceListening
    Dictate the shopkeeper's data…
    1 · TranscribeWhisper turns the audio into text, handling regional accents and store background noise. In this demo, the browser's native engine does the same job with no configuration needed.
    2 · Extract and assignA language model understands "eighteen twenty" ("mil ochocientos veinte"), "two thousand" ("dos lucas") or "five percent," identifies the product, and decides which field each figure belongs in.
    3 · ConfirmThe fields appear highlighted in green. The rep reviews with the shopkeeper and adjusts anything needed by hand before calculating.
    What it solves

    When perception crosses
    with the number, the decision appears.

    Perception×Real profit

    The shopkeeper believes the product that sells fastest is the one that earns them the most. With real purchase price, discount and sale price, the tool shows the monthly profit of each option, not gross sales.

    Their own numbers make the argument
    Combo×Isolated beverage

    The competitor's soda isn't sold on its own but bundled with the food item. The module separates food and beverage, and compares the current combo against the same combo with Postobón's Pareto SKU.

    Replacements ranked by profit
    Voice×Capture

    Eight data points per exercise, typed on screen in the middle of the store, cost time and attention. Dictated, they take under a minute, and the rep never looks away from the shopkeeper.

    Under a minute per product
    One SKU×The whole portfolio

    Once the first argument lands, the table of the 10 Pareto SKUs with the same assumptions shows how much the month would add up to by expanding the portfolio. Every row is editable.

    From swapping one product to expanding the order
    Survey×Close

    Four questions at the end: was it closed with the tool, what was achieved, how many boxes got committed? It connects the experience to the commercial outcome.

    Closing rate measurable by cluster
    Session×BI

    No manual report: every exercise emits a record with user, point of sale, duration, SKUs, data, result and survey. It's the raw material for the module 4 dashboard.

    Data ready for Business from day one
    Scope

    Five modules. Three already work
    in this demo.

    01
    Included

    SKU comparison

    Product selection, matching, capture, result and expanded portfolio.

    02
    Included

    Combo comparison

    Beverage/food breakdown, scenarios and replacement suggestion.

    03
    Included

    Survey and logging

    Visit close-out and automatic record, exportable as JSON or CSV.

    04
    Next phase

    BI dashboard

    Closing rate, most-combo'd SKUs, average duration and boxes committed per rep.

    05
    Preview

    Voice capture

    Already working in the demo with the browser's engine; in production with Whisper and an LLM.

    Responsive webapp down to 390 px, embeddable in WhatsApp Flows and Glüx · no backend in the prototype · price lists versioned by cluster in production