Skip to content
Demand Forecasting

How Growing D2C Brands Can Build Demand Planning Without an ERP

Your WMS exports and Excel sheets are enough to start. See how growing D2C brands build SKU-level forecasts, smarter replenishment and leaner inventory with an AI planning layer.

7 min readEnmovil Team

How Growing D2C Brands Can Build Demand Planning Without an ERP

Growing D2C brands move fast. A new product line, a marketplace listing, a quick commerce partnership and a festive campaign can all land in the same quarter. Each one adds demand signals, SKUs and warehouses to plan for. Yet many brands at this stage run operations on a WMS or OMS, a courier aggregator and a set of well-maintained spreadsheets, with an ERP rollout still on the roadmap.

Here is the good news: demand planning starts with data you already have, and an ERP is optional at the start. Sales history, stock positions, supplier lead times and your campaign calendar are enough to build a reliable, SKU-level demand plan. This guide shows how, drawing on the questions and priorities D2C operations teams raise most often in our conversations with them.

Why Demand Planning Gets Harder as D2C Brands Scale

In the early days, one person can hold the whole demand picture in their head. Growth changes that. These are the patterns D2C teams describe most often when they explain why spreadsheet planning is reaching its limit.

Demand now flows through six, seven or eight channels

A brand that began on its own website now sells through marketplaces, quick commerce apps, modern retail, corporate gifting and institutional orders. Each channel buys differently. Marketplace and quick commerce demand moves daily, while B2B and retail orders arrive in larger, lumpier batches. Planning them as one total number hides the patterns that matter.

SKU counts grow faster than the planning team

Variants, pack sizes and combo packs multiply quickly. Brands with a few dozen SKUs and brands with well over a thousand both tell us the same thing: the catalogue grows faster than the time available to plan each item well.

Planning lives in WMS exports and spreadsheets

A common workflow looks like this: download reports from the WMS or OMS, calculate a daily run rate, apply a fixed supplier lead time, compare against available stock and decide what to order. It works, and it takes a lot of manual effort every cycle. It also treats every day as an average day, which makes festive peaks and campaign spikes harder to prepare for.

Manufacturing, warehousing and transport are often outsourced

Many D2C brands work with contract manufacturers, third-party warehouses and local transporters. Outsourcing keeps operations lean, and it makes forward visibility even more valuable. Partners plan their own capacity, so a clear, early demand signal helps every partner deliver on time.

Shelf life turns forecast accuracy into margin

For food, wellness, personal care and nutrition brands, the right quantity at the right time keeps stock fresh. Accurate planning protects margin by keeping near-expiry stock and write-offs to a minimum.

Rethinking the ERP-First Approach

An ERP is a system of record. It captures transactions such as orders, invoices and stock movements. Demand planning needs a system of intelligence: something that reads those records, learns patterns and recommends what to do next. The two work best together, and the intelligence layer can arrive first.

A modern AI planning layer sits on top of whatever you run today. It connects through APIs where they exist, accepts Excel and CSV uploads where they are simpler, and can even pick up scheduled reports from a shared mailbox. When an ERP arrives later, the same planning layer connects to it, so the work you invest now carries forward.



Approach

What it means for a growing D2C brand

ERP first, planning later

Planning waits for a full implementation; the team continues with spreadsheets in the meantime

Planning layer first

Forecasts and replenishment recommendations start from existing WMS, OMS and Excel data; ERP connects in when ready

Planning layer on top of ERP

The ERP stays the system of record; the planning layer adds forecasting, optimisation and scenario planning

The Minimum Data You Need to Start

You do not need years of perfect data. You need a consistent view of a few core inputs. Most D2C brands already hold all of them.

Data input

Where it usually lives

What it powers

Sales history by SKU, channel and date

WMS / OMS, marketplace seller panels, website backend

Baseline demand forecast and seasonality

Current stock by warehouse

WMS, 3PL reports

Inventory position and transfer recommendations

Supplier lead times and MOQs

Purchase records, supplier sheets

Reorder points and order timing

Bill of materials (if you manage raw and packaging material)

Product or production sheets

Material requirement planning

Promotions and campaign calendar

Marketing plans, marketplace event calendars

Campaign uplift and what-if scenarios

Shelf life by SKU

Product master

Expiry-aware stocking and FEFO-led transfer

A 6-Step Framework to Build Demand Planning Without an ERP

Step 1: Bring your data into one planning layer

Connect your WMS, OMS, marketplace and website data through APIs where available. For everything else, use structured Excel or CSV templates on a set schedule. The goal is a single, refreshed view of demand and inventory, replacing the weekly ritual of stitching reports together.

Step 2: Clean and prepare the data automatically

Raw operational data carries returns, cancellations, stock-out days and one-off bulk orders. A good planning platform pre-processes this raw data into filtered data and shows you the difference between the two, so you can trust what the model learns from. It also flags whether the date range and volume are adequate before forecasting begins.

Step 3: Forecast at SKU × channel × warehouse level

Forecast where decisions are made. Machine learning and time-series models, combined with demand sensing, read recent order signals alongside regular-day run rates, festive peaks, campaign periods and seasonality. Forecasting by channel lets quick commerce, marketplace, D2C and B2B demand each follow their own rhythm.

Step 4: Turn the forecast into an inventory policy

A forecast becomes useful when it drives stock decisions. Use it to set dynamic safety stock and reorder points for every SKU at every warehouse, based on demand variability, supplier lead time and the service level you want to hold for each product class.

Step 5: Automate replenishment and procurement recommendations

With policies in place, the system recommends what to order, how much and when. It can propose supplier indents, suggest inter-warehouse transfers based on stock, demand velocity, expiry and proximity, and run MRP-style calculations from your bill of materials. Your team reviews and approves, spending its time on decisions instead of data assembly.

Step 6: Track accuracy and bias, and keep learning

Measure forecast accuracy and forecast bias every cycle, by SKU and channel. Compare predicted against actual demand, and let the model learn from each month of actuals. Bring sales, marketing and finance inputs into one monthly consensus plan, so campaigns and growth targets show up in the forecast before they show up in the warehouse.


Area

Spreadsheet-led planning

AI planning layer

Forecast basis

Average daily run rate

ML and time-series models with demand sensing, seasonality and campaign drivers

Granularity

Usually total or SKU level

SKU × channel × warehouse

Lead times

Fixed values per supplier

Actual turnaround tracked and applied dynamically

Safety stock

Static rule of thumb

Dynamic, based on variability and service targets

Scenario planning

New copy of the sheet for each scenario

What-if simulation for promotions, price changes and growth targets

Replenishment

Manual calculation each cycle

Recommended indents and inter-warehouse transfers for approval

Learning

Depends on individual memory

Accuracy and bias tracked; model refines with every cycle

Start Small: A Modular Roadmap for D2C Brands

The most practical path is modular. Begin with the capability that delivers value fastest, prove it, then expand.

1.      Phase 1: Demand forecasting. Upload sales data, run forecasts, compare them with your current spreadsheet output and review accuracy and bias dashboards.

2.     Phase 2: Inventory and replenishment planning. Add safety stock, reorder points, supplier lead times and replenishment recommendations across warehouses.

3.     Phase 3: Distribution and logistics planning. Extend into warehouse distribution, dispatch planning and shipment visibility as volumes and partners grow.

Making Adoption Easy for Lean Teams

Lean D2C teams need tools that fit into the day they already have. Three things make adoption smooth:

•      No specialist skill set required. The platform handles data preparation and model selection, so the team works with recommendations rather than algorithms.

•      Planning inside familiar tools. With an AI assistant such as Enmovil’s CADDIE, available in Microsoft Teams and Outlook, team members can ask for a forecast, an inventory view or a replenishment plan in plain language.

•      Human approval at every step. Recommendations come to the team for review, keeping experience and judgement at the centre of each decision.

How Enmovil Helps Growing D2C Brands Plan Demand

Enmovil’s Supply Chain Planning solution gives growing brands enterprise-grade demand forecasting, inventory planning and replenishment, built to work with the systems they already use. It connects with WMS, OMS, marketplaces and ERP systems through APIs, and accepts Excel uploads for everything else.

•      Demand forecasting and demand sensing at SKU, channel and warehouse level

•      Scenario modelling for promotions, price changes, campaigns and growth targets

•      Dynamic safety stock, reorder points and replenishment recommendations

•      Inter-warehouse transfer recommendations that factor in demand velocity, expiry and proximity

•      Forecast accuracy and bias dashboards that improve with every cycle

•      CADDIE, Enmovil’s AI assistant, for planning through prompts in Teams and Outlook

Brands can start with a single module and expand into logistics planning and execution as they scale.

Frequently Asked Questions

Can D2C brands do demand forecasting without an ERP?
Yes. D2C brands can forecast demand using sales data from their WMS, OMS, marketplaces and website, combined through API connections or Excel uploads. An AI planning layer cleans this data and produces SKU-level forecasts by channel and warehouse. An ERP can connect to the same layer later.
What data do I need to start demand planning?
Start with sales history by SKU, channel and date, current stock by warehouse, supplier lead times and your promotions calendar. Add a bill of materials if you manage raw and packaging materials, and shelf life if your products are perishable. More history helps, and the model improves as new data arrives.
How is demand sensing different from demand forecasting?
Demand forecasting estimates future demand from historical patterns, seasonality and planned events. Demand sensing adds recent signals, such as the latest orders, special days and campaign activity, to sharpen the near-term forecast. Together they give D2C brands a stable long-range plan and a responsive short-range view.
How should D2C brands plan inventory across marketplaces and quick commerce?
Forecast each channel separately, because marketplace, quick commerce, website and B2B demand move at different speeds. Then set safety stock and reorder points per warehouse based on each channel’s variability and the lead time to replenish it. Inter-warehouse transfers help rebalance stock as demand shifts.
Is an AI planning tool only for large enterprises?
Modern planning platforms are modular, so growing brands can begin with demand forecasting for their core SKUs and add inventory, replenishment and logistics capabilities over time. This keeps the starting scope focused and lets results from each phase guide the next investment.
How long does it take to get started?
Timelines depend on data readiness and the number of SKUs and channels. Brands starting with Excel-based forecasting can typically begin with a focused pilot on a subset of SKUs, review the output against current methods, and then expand the scope

Share this article

Your next step

Put CADDIE over what you already run.

Nothing ripped out. Leave your work email and we’ll show you CADDIE on your own operation, in 15 minutes.

Next: your name and company. That's all.

Air transport
Sea transport
Road transport