Name the limits. Let the tool pick the numbers.
Point the optimizer at the cells it may change, the cell you want as big or as small as possible, and the rules it must not break. It searches, and writes back the split that wins.
Method guide for Google Sheets Three solver engines free, up to 10,000 decision cells
Optimization in Google Sheets, three engines
Simplex LP is for models made of straight lines: budgets, blends, routes, rosters. It is exact and it is fast, and it handles integer and binary decision variables, so “how many trucks” and “which sites do we open” are both fair questions.
Nonlinear takes over when the curve bends, which is what happens the moment a channel has diminishing returns or a price moves the volume it sells. Evolutionary is the one for models full of IF and LOOKUP, where the sheet is not smooth enough for either of the others to get a grip.
The report names the constraints that are actually binding your answer, which is usually the more useful half. Knowing the plan is capped by one warehouse rather than by the budget is what changes next quarter.
Which tool, and what it is for
One panel, one Solving method dropdown. If a model is not linear, Simplex says so rather than guessing, and you switch.
The three solving methods
- Simplex LP straight-line models. Exact, fast, and takes integer and binary variables.
- Nonlinear smooth curves: diminishing returns, price and volume, compounding.
- Evolutionary rough or non-smooth models, the ones built from IF and LOOKUP.
What you can ask for
- Maximize or minimize a cell profit, cost, miles, headcount, waste.
- Hit a target value on the Simplex and Nonlinear engines.
- Constraints, in your own cells floors, ceilings, totals that must balance, integers, binaries.
When the inputs themselves are uncertain
- Optimization under uncertainty it simulates every candidate rather than assuming one number. A Pro engine, on the optimization under uncertainty page.
The answer, and what is holding it back
A number on its own does not survive the meeting. The report gives the status, the objective it reached, the value of every decision cell, and which constraints are binding, so you can argue with the limit instead of with the arithmetic.
It writes into a tab of your own spreadsheet, next to the model it solved. Nothing is exported, and nothing has to be re-imported.
Free or Pro
All three methods are free on every plan, and their size limit is the method, not the license. Simplex LP answers a continuous model of up to 2,000 decision cells and 600 constraints in about a second, and it does not stop there: up to 5,000 cells and 1,200 constraints run on the same engine with a progress line and a Cancel button, and up to 10,000 cells and 3,000 constraints run on a sparse engine built for models where each constraint touches only a few cells. Evolutionary takes up to 100 cells, Nonlinear up to 60. The panel counts the cells you list under Cells Sortia may change and tells you the number before you run. There is no run counter, and Pro raises none of these: we charge for the Pro tools, never for the size of your model.
Two things about the big end of that ladder, because a ceiling stated without them is a half-truth. The sparse engine is for continuous models: ask Simplex LP for whole numbers or yes/no decisions and it proves the optimum for up to 500 of them, inside the 2,000-cell instant ceiling, because the method that proves an integer answer cannot use the sparse path at all. Evolutionary takes 30 whole numbers and cannot prove one. And how long the big end takes depends on how coupled your model is, not only on its size: a 10,000-cell model whose constraints each touch a few cells was measured at 22.7 seconds, while the same size densely coupled takes minutes. The panel says which it expects before it runs. Every answer the sparse engine returns is re-checked by an independent verifier before it can reach your sheet, so a faster path is not a path you have to take on trust. The measured ladder is on the performance page.
Those runs happen in your browser by default. A model the browser cannot read falls back to solving through the sheet itself, recalculating your workbook at every step, and that path is slower and its limits are lower: 200 decision cells on Simplex LP, 15 on Evolutionary, 10 on Nonlinear. The panel names the reason it fell back before it writes anything.
The Pro cousin is optimization under uncertainty, where every candidate is simulated rather than assumed. Every install gets five full-quality runs, shared across all five Pro engines rather than five for each. Pro is $199/year.
Worked examples to start from
Each one opens with the decision cells, the objective and the constraints already set, so you can press Run and then change one rule to see what moves.
- A fixed budget split across channels with diminishing returnsSplit the Ad Budget for Maximum Orders
- The order of stops that drives the fewest milesShortest Route Finder
- Who should work on which projectBest-Fit Assignments
The other 39 models in the library that use one of these:
- What is the cheapest meal plan that hits your targets?Cheapest Meal Plan (Optimization)
- How few nurses can cover every hour of your unit?How Many Nurses Does Each Shift Need?
- Which offer package should you actually ask for?Job Offer Package Optimizer
- How good is each team, really?Power Ratings & Spread Predictor
- How much should you produce each month?Production & Demand Planner
- Which projects should we fund this year?Project Portfolio Selector
- Where should your warehouse actually be?Warehouse Location Optimizer
- Should we pay the overtime or take the late penalty?Overtime vs Late Penalty Scheduler
- Which mix of stores should fill your square footage?Retail Space Mix Optimizer
- How do you give every team a fair share of every skill?Balanced Team Builder
- How should you actually spend your 168 hours a week?Design Your 168-Hour Week
- Is there a deal that beats splitting the difference?Win-Win Deal Finder
- Which plant should ship to which store?Which Plant Ships to Which Store?
- How many boards do you actually have to buy?The Fewest Boards for the Cut List
- Who is actually carrying the month?Which Lawyer Takes Which Matter?
- Cheapest crew per site, or cheapest week?Which Crew Goes to Which Site?
- Everyone fits. Did anyone actually get what they wanted?Seat Everyone Without Breaking the Rules
- What is the cheapest menu that still keeps every promise?Feed Everyone for the Least Money
- Are you paying for cement the spec does not need?Which Concrete Mix Meets Spec for the Least Money?
- Is the biggest engagement worth the two it blocks?Which Engagements Should We Take?
- Is filling every billable hour the wrong plan?Where Should My Own Week Go?
- Packing the best thing first. What does that cost you?Fit the Most Into One Box
- How many people does the rota actually need?Cover Every Hour With the Fewest People
- Is the plan limited by money or by people?Split a Fixed Budget Across Three Things
- Does the cheapest quote stay cheapest after freight?Which Suppliers Fill the Order Cheapest?
- Is an equal share of the work actually fair?Share the Marking Fairly
- Commit to the servers or rent them by the hour?Which Instance Mix Holds the Load for the Least Money?
- Are you paying for machine days you never use?Which Plant Do We Hire and Which Do We Own?
- Is it cheaper to staff the night or to buy it in?Cover the Support Hours for the Least Money
- Which class should go to the hired hall?Fit the Timetable Without a Clash
- What is the standing recipe costing you a tonne?The Cheapest Blend That Meets the Spec
- What is your diversification rule costing you?Which Deals Can We Actually Fund?
- What does one conflict of interest cost you?Who Works on Which Client?
- What does your target return cost you in risk?The Cheapest Way to Hit the Target Return
- Which client should you turn down?Which Retainers Fit the Team We Have?
- Is the round your driver does now the shortest?Which Order of Stops Makes the Shortest Delivery Round?
- Does your roadmap leave revenue on the floor?Which Features Fit the Quarter?
- Should you make more of your most profitable product?What Should We Make This Week?
- How few people can cover a seven-day week?How Few People Can Staff the Week?
Try it in your own sheet
- Open Sortia in Google Sheets and choose Start from a template.
- Pick one of the models above, and it loads with the inputs filled in.
- Change the assumptions to fit your situation and press Run.
Never used Google Sheets? Start here goes the whole way, in seven steps, and assumes nothing.
Other methods: Monte Carlo Decision trees Schedule risk Critical chain Optimization under uncertainty Statistics Machine learning Forecasting What-if analysis