65% lower
product-level forecast error
Error fell from 57% to 20% across more than 7,200 products in a multi-location retail operation.
Demand & inventory planning
Know what to buy, how much, and when.
ToolPlex builds and runs planning systems for retailers, distributors, manufacturers, and other multi-product businesses. Forecasts are combined with current stock, open orders, lead times, and business rules to produce purchasing plans, replenishment recommendations, and stock warnings.
Purchasing plansReplenishment and allocationImport and seasonal planningExcess-stock guidance
Selected results
A few results from our work.
34% lower
forecast error in a retail trial
Weekly store-and-product forecasts were compared with the retailer’s existing planning method.
20–32% lower
error for purchase planning
A multi-branch distributor received a five-month import plan alongside weekly reorder and excess-stock guidance.
The problem
A forecast should help you decide what to buy.
An ERP report or spreadsheet may estimate demand, but buyers still have to account for current stock, lead times, open orders, and purchasing constraints.
The same rule is used for every product
Fast movers, intermittent items, seasonal lines, and new products behave differently, but the planning sheet treats them alike.
Planners override the output constantly
Experienced staff replace the system’s number using recent sales, upcoming events, and several other reports.
Stockouts make low sales look like low demand
Low sales can mean low demand or simply that stock was unavailable. Treating both cases the same produces a poor forecast.
Accuracy is measured at the wrong level
A national monthly score can look good even when forecasts are poor for the specific stores and products buyers must plan.
What ToolPlex builds
Get forecasts, purchasing plans, stock warnings, and accuracy reports in one place.
We connect the source data, generate forecasts, turn them into purchasing recommendations, and monitor accuracy after launch.
Compare with the current method
We first measure the method your team already uses for the same products, locations, and planning period.
Choose a method that fits the data
We compare statistical methods, custom machine-learning models, or a combination, then use the simplest approach that performs reliably.
Turn forecasts into purchasing plans
We use your lead times, stock targets, case sizes, and open orders to turn the forecast into practical purchasing recommendations.
Clear results for planners
Teams receive dashboards or files formatted for their existing systems, with clear comparisons against recent demand.
Checks before planners receive it
We do not publish a forecast if it performs worse than the current method or fails an agreed data check.
Maintenance after launch
ToolPlex maintains the data feeds, scheduled runs, accuracy checks, and updates needed as products and processes change.
Forecasting itself uses statistical or machine-learning methods. Large language models can help buyers investigate a recommendation and find supporting information.
Connects with the software and files you already use:
ERP and purchasing systemsPOS and sales historyWarehouse and inventory dataOpen orders and quotesSpreadsheets and flat filesProduct and store masters
How we build it
From the first review to launch.
- 01
Understand the purchasing decision
We learn what buyers decide, how far ahead they plan, which limits they work within, and where the current method falls short.
- 02
Test against the current method
We clean the history and compare each approach with the current method using past periods held back for testing.
- 03
Deliver it in a useful format
Buyers receive the forecast through ToolPlex, an export, or the planning system they already use.
- 04
Monitor and maintain it
We track accuracy, missing data, planner overrides, and business changes, then update the system when needed.
Philippines
Demand forecasting for Philippine operations
ToolPlex began this work with Philippine retailers, distributors, and manufacturers. Planning teams here often deal with seasonal demand, long import lead times, fragmented store data, new-product launches, and purchasing decisions that still depend on spreadsheets or individual judgment.
- Forecasts can reflect Philippine holiday and seasonal patterns when the data supports them.
- Forecasts can be delivered in the ERP, JDA, purchasing, or spreadsheet formats the team already uses.
The business case
The cost goes beyond staff time.
It may show up as delayed decisions, excess stock, missed sales, or unreliable reports. We start with one area where the improvement can be checked.
A good fit for ToolPlex
- Purchasing or inventory decisions repeat often enough to measure
- The business carries many products, locations, or seasonal lines
- Planners spend substantial time correcting the current method
How we reduce the risk
- Begin with one recurring job
- Compare the result with the current method
- Add more only after the first part works
FAQ
Frequently asked questions.
Do we need clean data before starting?
No. Reviewing and cleaning the data is normally part of the project. We do need enough reliable history to show whether a new forecast is better than the current method. If the data is not sufficient, we will say so.
Does ToolPlex replace our ERP or planning system?
Usually not. ToolPlex connects to the systems already in use and can return forecasts or purchasing outputs in the format those systems expect.
What forecasting model do you use?
There is no single model for every company. We compare statistical and machine-learning approaches with the client’s current method, then put the simplest reliable one into production for the period the team needs to plan.
Can you forecast new products or seasonal demand?
Yes, when the business has similar products, useful product attributes, launch history, or repeated seasonal patterns. We test these forecasts separately because new products are less predictable than established ones.
Show us the work you want to improve.
We'll explain what we can build, how it would connect, and where we would start.