WonForge changes the equation

We deliver end-to-end supply chain planning built specifically for the physical realities of the Process Industry. We optimize your most difficult sequencing logic (like complex grade-to-grade transitions) to minimize changeover waste, while strictly enforcing your inventory and shipment constraints. We stripped away the historic barriers to entry so you only pay for the value:

• Zero IT Burden: 100% web-based and operating on a fully managed AWS infrastructure.
• Turnkey Architecture: We manage the security, the maintenance, and the commercial solver licensing.
• Shared Value: You get custom optimization models and true Decision Intelligence without the enterprise price tag.

Key Point: You shouldn't have to choose between a fragile spreadsheet and an overpriced IT nightmare.

Safety Buffers on Top of Safety Buffers

Most plants operate with massive invisible layers of protection because static planning tools cannot calculate the "ripple effect" of a production change in real-time.

• The Spreadsheet Survival Strategy: Research shows that over 80% of supply chain planning processes still rely on Excel. Because these sheets are disconnected from the true physics of the plant floor, planners are forced to maintain a 20% "fear buffer" to handle the unexpected.
• The Operational Safety Net: To protect against high-level alarms or spill risks during complex changeovers, floor operators often keep another 10% empty "just in case."
• The "Hidden Tax" on OEE: This cumulative 30% buffer is a direct tax on your margins. It forces shorter campaigns, more frequent washouts, and unnecessary waste because the system "thinks" the plant is full when it's actually just poorly synchronized.

Key Point: Ghost Capacity is the gap between what your site map says you have and what your operation dares to use—and it's costing you every day.

The Barrier is the Tool, Not the Talent

Planners in this niche are master navigators of "Tank Tetris," but they are often handcuffed by ERP outputs that use static planning logic. Most ERPs assume infinite capacity and ignore the reality of sequence-dependent changeovers—like the fact that switching from a high-viscosity resin to a light coating requires a 6-hour caustic wash. When the software doesn't understand the physics, the only way a planner can protect the plant is with a safety margin.

Key Point: The constraint isn't your people; it's that the tools don't model the physics, so the only defense is buffer.

The WonForge Value: Empowering the "Strategic" Planner

WonForge provides the Decision Intelligence Engine that turns these operational "fear buffers" into a profit center.

• Synchronized Confidence: Our engine simulates thousands of scenarios across your entire network, proving exactly how much head-space is needed. When the math guarantees a tank can be safely filled to 98%, that "Ghost Capacity" suddenly becomes real, usable volume.
• Unlocking Hidden Capacity: Instead of spending $100,000+ on a new 50,000-gallon tank installation, you can reclaim 15-20% of the capacity you already own.
• Strategic Leverage: Your team can finally move from "making it fit" to "making it profitable"—using that reclaimed capacity to buy raw materials at lower prices or pre-build inventory for peak season.

Key Point: Turn fear buffers into a profit center with high-fidelity tools that unlock the full potential of your existing assets.

1. "First Feasible" vs. "Most Profitable"

A spreadsheet can't evaluate margin per gallon across your full schedule. If Generic Product A is running tomorrow and allows 10% rework, the sheet (or the rule you encode in it) says use it to clear the tank. Product A uses cheap base stocks—replacing them saves you pennies. The optimized choice: The engine sees Premium Product B scheduled for Friday, which uses expensive synthetic raw materials. It calculates that holding the inventory for 3 days captures 4x the savings per gallon, justifying the storage cost.

• The Problem: Product A uses cheap base stocks. Replacing them saves you pennies.
• The Optimized Choice: Premium Product B uses expensive synthetic raw materials. Holding for 3 days captures 4x the savings per gallon, justifying the storage cost.

Key Point: The engine holds rework for high-margin formulations when the math justifies it—turning storage from a problem into strategic leverage.

2. Finding the "Long Tail" Opportunities

Spreadsheets typically rely on a short list—a "Top 5" of formulations that allow rework—because scanning the full formulation library in Excel is impractical. But your technical database might contain 50 low-volume SKUs that are also valid candidates. WonForge scans the entire formulation library instantly. It identifies consumption opportunities in the "long tail" of your portfolio—obscure batches that relieve tank pressure without impacting your primary products.

Key Point: WonForge scans the entire formulation library instantly, identifying long-tail opportunities that spreadsheets can't easily surface.

3. Storage as Strategic Leverage (The Shadow Price)

We don't use static safety rules like "Sell if > 90% full." Optimization treats tank capacity as a dynamic economic constraint. The engine calculates the precise value of your storage capacity dynamically. When tank pressure is low, the engine is "selective," holding out for high-margin formulations. As tank pressure rises, the "cost" of holding inventory increases mathematically, forcing the model to accept lower-margin options to protect the sequence.

Key Point: It's a continuous sliding scale, not a rigid rule—storage is valued dynamically so every gallon goes to its most profitable home.

1. The Discrete Bias

In assembly lines (Discrete), inventory between stations is waste. It hides defects. Consultants apply this logic to process plants, not realizing that in liquids, Inventory equals Time.

Key Point: They see idle cash. The Plant Manager sees the only thing keeping the plant running.

2. The War of Two Physics

You are managing a conflict between Upstream and Downstream. The Reactor (Upstream) wants to run continuous 24/7. The Packaging Line (Downstream) is stop-and-start.

• Without a buffer, every time a labeler jams downstream, the reactor must stop.
• In process manufacturing, 'stopping' isn't just lost time—it can mean material hardening, burn-in, or full cleaning cycles.

Key Point: The liquid in the tank is the mechanical shock absorber that decouples these two conflicting physics.

3. The Cost of Cutting Too Deep

When you arbitrarily cut tank levels to hit a financial KPI, you remove the shock absorber. You might save $50k in working capital, but you lose $500k in OEE and Yield because your reactor keeps tripping.

Key Point: Don't minimize inventory. Optimize the buffer.

1. The Amazon vs. The Railcar

Standard inventory models (like EOQ) calculate optimal production quantities based on holding costs and demand. This works for packaged goods because the container is disposable. It fails for bulk because the container is a capital asset.

• The Amazon Model (Packaged Goods): You optimize for the item inside the box. Once the truck leaves the dock, the logistics transaction is over. The constraint is the inventory on the shelf.
• The Chemical Reality (Bulk Goods): You must optimize for the "box" itself. Unlike a cardboard box, your railcar operates in a closed loop. If you ship product today, you decrement your available loading capacity for the next 45 days.

Key Point: While both industries deal with lot sizes, only the bulk industry has its production capacity dictated by the return trip of its packaging. If your fleet is stuck at a customer site, your plant stops—even if you have empty tanks and full silos.

2. The Hidden "Return Loop" Trap

The most common failure in chemical planning isn't running out of raw materials; it is running out of empty cars to fill. This is the concept of the "Rolling Warehouse." Your inventory isn't just sitting in a static tank at the plant; it is strung out across 2,000 miles of track.

• The Surge: Sales closes a massive deal. Production ramps up. You ship 50 cars in Week 1.
• The Crash: In Week 4, those cars are still rolling back empty. Your plant has the product to make, but the loading track is empty. You are forced to idle the plant.

Key Point: If you don't plan the return trip (RTT), you are planning a shutdown.

3. The Fix: Justifying the Investment

Most schedulers treat the fleet size as a fixed wall: "We only have 100 cars, so we have to cut sales." This is the wrong approach. The goal of optimization isn't just to restrict your plan to fit the fleet; it is to identify when the fleet is the bottleneck so you can fix it.

• WonForge models the fleet constraint to calculate the Shadow Price of every missing railcar.
• We don't just tell you that you are short; we provide the data to prove that leasing 5 additional cars will unlock $200k in monthly profit.

Key Point: You move from "managing shortages" to "justifying equipment investment" with hard math.

1. The Physics of the Matrix

In discrete manufacturing (like assembling cars), a "setup" is usually about changing tools. In process manufacturing (chemicals, dairy, food), a "setup" is about physics and chemistry. The Clean-In-Place (CIP) process is governed by a matrix of constraints that most planning tools simply ignore:

• Color Sequencing: Light → Dark is cheap. Dark → Light is expensive.
• Allergen Breaks: Running a non-allergen after an allergen requires a "Full Wash" that halts the line for hours.
• Viscosity & Chemistry: Some products react with each other. You can't just follow Product A with Product B; you might need an intermediate "buffer" product or a solvent cycle.

Key Point: Your operators are likely manually overriding the schedule every week to fix these sequencing errors. They know that following the "official" plan would result in unnecessary 6-hour washes, so they reshuffle the deck based on tribal knowledge.

2. The Combinatorial Trap

Why can't a human planner just "group like with like" to minimize washing? It works fine if you have one tank and three products. But that isn't your reality. You likely have dozens of tanks, shared headers, manifold constraints, and hundreds of recipes. This turns the scheduling problem into a variation of the "Traveling Salesman Problem."

• If you have 5 tanks and 20 orders, there are millions of possible sequences.
• A human brain—or a spreadsheet—cannot calculate the total "CIP Penalty" for every possible combination to find the global minimum.
• Humans default to "Greedy Optimization": they pick the best next move. But taking a short changeover now might force a catastrophic 12-hour washout later in the week.

Key Point: You don't have a scheduling problem. You have a mathematical optimization problem that exceeds human cognitive capacity.

3. The Hidden Cost of Sustainability

Optimizing your CIP matrix isn't just about unlocking hidden capacity (though it often yields 5-10% more throughput). It is a direct financial lever for your variable costs. Every unnecessary "Full Wash" consumes:

• Water: Thousands of gallons of high-quality industrial water.
• Energy: The BTU cost of heating that water to 180°F (82°C).
• Chemicals: Expensive caustic sodas and sanitizers that erode your margin.
• Effluent: The cost to treat the wastewater you just generated.

Key Point: When you optimize the sequence to minimize the severity of washes, you are attacking your utility bill and your carbon footprint simultaneously.

1. The "Average Cost" Trap

Most manufacturers calculate margin using "Standard Costing." They take total plant overhead and divide it by total volume. This smears the cost of complexity across everyone.

• The Reality: The customer buying the standard white paint is subsidizing the customer buying the small-batch neon green paint.
• The Problem: You think both customers are profitable. In reality, one is a cash cow, and the other is a cash drain.

Key Point: To see the truth, you need to stop looking at averages and start looking at Scenarios.

2. The Power of Parallel Scenarios

It is impossible to find the true cost of a complex customer simply by looking at a spreadsheet. The only way to measure their actual impact is to run the numbers two ways: with them and without them. This is where WonForge's Scenario Planning steps in. You can clone your production model and solve two parallel futures instantly:

• Scenario A (Baseline): The optimal plan including the complex customer's orders.
• Scenario B (Alternative): The optimal plan excluding or deferring those orders.

Key Point: By comparing the bottom line of these two scenarios, the "hidden costs" immediately surface. You might find that removing $50k of revenue from a difficult customer actually increases your total profit by eliminating costly changeovers and unlocking capacity for higher-margin goods.

3. From "Can We Make It?" to "Should We Make It?"

Most planning systems ask, "Do we have the capacity to fill this order?" Scenario-based optimization asks, "Does filling this order increase our total EBITDA?" You can test strategic questions before you sign the contract:

• What-If: We accept this rush order? (Does it force a break in a high-efficiency campaign?)
• What-If: We enforce a strict "Full Tank" Minimum Order Quantity (MOQ)? (How much waste does that eliminate?)
• What-If: We shift this customer's volume to next week? (Does that smooth out the labor requirement?)

Key Point: Scenario-based optimization shifts planning from feasibility to profitability.

4. Empowering the Commercial Conversation

This doesn't mean you have to fire the customer. Once you run these scenarios, you enter negotiations with hard evidence.

• Strategic Surcharges: "We ran a simulation, and your custom formulation triggers a cleaning cycle that costs us $4,000 per run. We need to adjust the price to cover that impact."
• Win-Win Adjustments: "Our model shows that if you can accept delivery on Thursdays instead of Mondays, we can reduce your price because it fits our optimal sequence."

Key Point: Armed with scenario analysis, you can turn unprofitable relationships into win-win partnerships through data-driven negotiations.

1. The "Siloed Safety Stock" Trap

Most organizations calculate safety stock using standard textbook formulas applied to each location independently. The Plant ensures they have enough stock to feed the Central DC. The Central DC buffers stock to feed the Regional Warehouses. The Regional Warehouse buffers stock to service the customer.

Key Point: When every node buffers against the variability of the node before it, you create a static version of the bullwhip effect. Instead of amplifying demand signals, you are amplifying inventory buffers. You are essentially paying to insure the same risk multiple times across the network. Crucially, this approach also obscures real risks; while one warehouse sits on excess stock, another might be starving, leading to stockouts despite high total network inventory.

2. Moving from Local to Network Optimization

Multi-Echelon Inventory Optimization (MEIO) changes the math. Instead of looking at one warehouse at a time, the WonForge engine looks at the entire chain simultaneously. It asks: "Where is the cheapest and most effective place to hold this safety stock to guarantee our service targets?" By shifting the perspective, the engine identifies opportunities to simultaneously reduce inventory and improve availability that traditional planning methods miss:

• Risk Pooling: If the Central DC is highly reliable, the Regional Warehouses can carry significantly less safety stock without risking stockouts.
• Postponement: It may be optimal to keep inventory in a semi-finished (bulk) state longer, only packaging it when demand is confirmed. This reduces the risk of having too much of the wrong SKU and not enough of the right one.
• Lead Time Sensitivity: It identifies which nodes actually require rapid response versus which can tolerate replenishment delays.

Key Point: The engine identifies opportunities to simultaneously reduce inventory and improve availability.

3. The Perishability Factor

For WonForge clients in the food & beverage and chemical sectors, excess inventory isn't just a cash flow problem—it's a waste problem. Holding high safety stock in a siloed manner increases the risk of product expiring before it reaches the customer.

Key Point: WonForge's decision engine incorporates shelf-life constraints directly into the inventory model. We calculate the "freshness risk" of every stocking decision, ensuring that aggressive safety stock targets don't result in aggressive write-offs.

4. The Financial Impact

By transitioning from sequential planning to multi-echelon optimization, companies typically see a reduction in total network inventory of 15–30%—often while actually improving on-shelf availability. In a high-volume chemical business, this often represents enough released cash to fund a new production line or cover the entire R&D budget for the year.

Key Point: Multi-echelon optimization releases working capital that can be reinvested in growth initiatives while maintaining or improving service levels.

The "Tank Tetris" Problem

Consider a common scenario: You have a rush order for Product A. Your planner looks at the schedule: Reactor 1 is open. The raw materials are available. The changeover is minimal. A human planner greenlights the run. It looks like an efficiency win. But this decision can silently destroy plant capacity later in the week.

The Invisible Domino Effect

Here is what the human planner couldn't see (because the calculation is too complex to simulate mentally):

• The Constraint: Product A must go into Storage Tank 4 (the only heated tank available).
• The Conflict: Product B, currently running on Reactor 2, is scheduled to finish tomorrow. It also requires a heated tank.
• The Block: Because Tank 4 is now occupied by the rush order, Product B has nowhere to go.
• The Ripple: Reactor 2 must hold the batch (blocking new production) or dump it. Consequently, it misses its window to feed Packaging Line 3, causing a gap in Finished Goods production two days later.

Key Point: The planner "saved" the rush order but unwittingly sacrificed total plant throughput.

How the WonForge Engine Handles "The Ripple"

While a human planner looks for an open slot, the WonForge engine simulates the entire financial and operational flow. In this same scenario, the engine doesn't simply say "yes" or "no." It weighs the total economic impact:

• Future-State Visibility: The engine projects inventory positions across the planning horizon. It immediately identifies that filling the tank on Day 1 creates a blockage on Day 2.
• Trade-off Analysis: It calculates the margin impact. Is the rush order worth enough to justify blocking Reactor 2 and starving the packaging line?
• The Optimized Path: It evaluates thousands of moves to find the highest-profit outcome. It might suggest delaying Product A until a tank clears, or prioritizing Packaging Line 3 to drain the tank early.

Key Point: It doesn't rely on hope. It ensures the decision is based on math, not a blind spot.

The Bottom Line

We aren't replacing your planner's expertise; we are elevating it. While a human focuses on specific units or days, WonForge solves the entire operation holistically. It treats your reactors, tanks, and packaging lines as a single, unified system, ensuring that local efficiencies never come at the cost of global profitability.

The Hidden Cost of Suboptimal Production Planning

Every manufacturer has a recipe defining what goes into each product. But in process industries like chemicals, coatings, and specialty materials, a recipe is more than an ingredient list. It's the digital fingerprint of how your plant actually runs — how raw materials flow, what by-products emerge, and how campaigns are sequenced to balance throughput, cost, and yield.

Key Point: Every suboptimal production decision directly impacts EBITDA and margin performance.

How Chemical Manufacturing Optimization Delivers 12-20% Asset Utilization Gains

Traditional ERP and spreadsheet-based planning tools were built around static assumptions: fixed recipes, fixed yields, and fixed routings. But modern process manufacturing is dynamic. Each batch or campaign interacts with real-world variation — material properties, capacity limits, and market shifts.

• Recipes with fixed input/output ratios that ignore yield variability
• Potency and concentration handled via manual adjustments
• By-products tracked separately from cost and value impact
• Campaigns scheduled line-by-line, not across shared resources
• No holistic trade-off between cost, throughput, and service

Key Point: Plans that satisfy production constraints but miss margin opportunities — or create inefficiencies hidden in day-to-day operations.

What is Recipe Optimization in Chemical Manufacturing?

Recipe optimization in chemical manufacturing is the process of using mathematical optimization solvers to automatically evaluate every aspect of production planning — demand patterns, recipes, routings, campaigns, yields, and constraints — to find the most profitable plan given actual business conditions.

Key Point: The solver doesn't just find a way to make everything fit — it balances economics, throughput, and customer demand simultaneously, producing a plan that is both executable and financially optimal.

The Business Case: Typical Results from Chemical Manufacturing Optimization

Chemical companies implementing optimization-driven recipe modeling report significant improvements across key metrics:

• 12-20% improvement in asset utilization through smarter campaign sequencing
• 8-15% reduction in working capital from reduced intermediate inventory buildup
• 6-12% margin improvement through optimal routing decisions
• 50-70% reduction in planning cycle time — from weeks to hours

Key Point: These improvements translate directly to measurable margin gains and improved financial performance.

How WonForge Captures True Operational Complexity

WonForge is built around a flexible decision engine designed for complex, process-oriented manufacturing environments. Its strength lies in representing how recipes, campaigns, and capacities actually behave — not how they're simplified in spreadsheets.

Key Point: The result is a dynamic operational model of your manufacturing network — one that adapts to changes in demand, feedstock, or capacity and automatically recalculates the optimal plan.

1. Architecting the Flexible Network

Resilience isn't an accident; it must be designed into the system. Optimization models allow planners to incorporate real-world trade-offs and constraints, ensuring your plans remain viable even when things go sideways.

• De-risking Single Sourcing: Evaluating the actual cost/benefit of multi-sourcing raw materials to reduce critical supplier dependency (a necessity when dealing with volatile inputs).
• Flexible Production Changeovers: Modeling flexible production lines that can quickly switch between recipes or product families, minimizing lost time.
• Strategic Buffer Stock: Optimizing the location and type of buffer inventory—not just how much—to shield final customers from upstream delays (crucial for short shelf-life items).
• Alternative Routes: Pre-planning alternative transportation methods and routes to bypass logistics bottlenecks.

Key Point: Optimization models allow planners to incorporate real-world trade-offs and constraints, ensuring your plans remain viable even when things go sideways.

2. The Power of Real-Time Response

When a major disruption hits—be it a facility breakdown or a sudden spike in demand—the critical difference is the speed of your replan. Automated optimization systems can run thousands of scenarios and generate a robust, executable plan in minutes, not days.

• Instant Capacity Reallocation: Rapidly re-routing high-priority orders to alternative lines when a production asset fails, automatically balancing cost and service level.
• Dynamic Routing & Delivery: Generating instantaneous adjustments to trucking or vessel routes when port or traffic disruptions occur.
• Inventory Rebalancing: Quickly evaluating and moving the right inventory (especially specialized inputs) across distribution centers to meet revised demand or production needs.

Key Point: This ability to run rapid calculations is where modern solutions using high-speed solvers truly separate themselves from legacy systems.

3. Rigorous Scenario Analysis and Stress Testing

Optimization is the ultimate "What-If" engine. It allows your team to rigorously stress-test the supply chain before a crisis, turning hidden risks into defined contingency plans.

• Supplier Failure Modeling: Quantifying the total impact of a top-tier supplier going down and validating the effectiveness of your pre-defined backup plans.
• Demand Volatility Scenarios: Stress-testing capacity against severe demand surge or collapse and identifying the labor and inventory limits under those conditions.
• Cost of Resilience: Evaluating the precise cost premium (e.g., in inventory or CapEx) required to achieve a target level of operational resilience.

Key Point: Optimization allows your team to rigorously stress-test the supply chain before a crisis, turning hidden risks into defined contingency plans.

4. Continuous Adaptiveness and Learning

Long-term resilience requires a system that learns and evolves. An effective optimization model is never static; it's constantly improving its own constraints and inputs.

• Forecasting Integration: Seamlessly integrating machine learning outputs for better demand and lead-time forecasts directly into the optimization constraints.
• Adaptive Constraints: Automatically adjusting model parameters based on continuous real-world performance feedback (e.g., tightening the safety stock requirement when supplier performance has been excellent).
• Operationalizing Learnings: Using every disruption event to immediately update and improve the model's structure for the next run.

Key Point: An effective optimization model is never static; it's constantly improving its own constraints and inputs.

1. Spreadsheets Manage Data, Not Decisions

Spreadsheets can store production rates, inventory targets, and forecasts — but they can't decide how to balance them. Once the number of products, lines, and constraints grows, manual adjustments turn into guesswork.

Key Point: Optimization models evaluate millions of possibilities instantly to find the best plan under all constraints — something no spreadsheet can achieve.

2. Real-World Constraints Don't Fit in a Cell

Manufacturing isn't just math. Real planning involves setup times, shared equipment, batch rules, and material dependencies — all changing daily. Spreadsheets rely on human memory and fragile formulas to maintain these relationships. One broken formula can derail an entire week's plan.

Key Point: Optimization models handle these constraints natively, ensuring every plan is both feasible and efficient.

3. No Easy Way to See Trade-Offs

Good planning is about trade-offs. A planner might ask:

• What if demand spikes by 20%?
• What if a line goes down?
• What if we run larger batches to reduce setups?

Key Point: In spreadsheets, exploring each "what-if" means copying and rebuilding the model. Optimization models simulate these scenarios instantly, showing how every decision affects cost, utilization, and service — empowering planners to make faster, smarter decisions.

4. Static Tools in a Dynamic World

Production data changes constantly — new orders, machine downtime, material delays. Spreadsheets can't adapt in real time. By the time one version is finalized, the situation has already changed.

Key Point: Automated optimization models can re-optimize multiple times per day, integrating live data and generating updated schedules in minutes.

5. The Hidden Cost of Errors

Even the most carefully built spreadsheets contain mistakes. Studies show over 80% of complex spreadsheets include at least one serious error. In manufacturing, those errors lead to costly consequences — excess inventory, missed shipments, and unnecessary overtime.

Key Point: Optimization models eliminate this risk by replacing manual formulas with consistent, traceable, and auditable logic.

6. Time and Talent: The Real Competitive Advantage

Traditional planning cycles often take days or even weeks to finalize. That delay creates missed opportunities, inefficient resource use, and higher operating costs. Meanwhile, manual planning burns valuable talent — planners spend most of their time fixing data, reconciling files, and updating spreadsheets instead of analyzing and improving operations.

Key Point: Optimization models compress the planning cycle from days to hours, allowing planners to react immediately to changing conditions. More importantly, automation frees them from repetitive data work, letting them focus on strategy, risk management, and process improvement — the work that actually drives business performance.

1. Machine Learning: Learning from Patterns

Machine learning is best suited for prediction — discovering relationships hidden in historical data.

• What will demand be next week?
• How long will a shipment take to arrive?
• What's the likelihood of a supplier delay?

Key Point: ML finds patterns in data — but it doesn't decide what to do next. It tells you what's likely, not what's optimal.

2. Optimization: Making the Best Possible Decision

Optimization is about prescription, not prediction. It takes data (including ML forecasts) and identifies the best decisions under constraints.

• How to allocate limited capacity across product lines.
• Which production schedule minimizes cost while meeting service targets.
• How to assign workforce shifts to balance productivity and coverage.

Key Point: Optimization handles trade-offs and finds the best solution under constraints — something ML cannot do.

3. How They Work Together

In real systems, the two often operate in sequence:

Key Point: Together, they turn uncertainty into action: ML provides insight; optimization delivers the plan.

4. The Takeaway

If the goal is to understand or forecast behavior, use machine learning. If the goal is to make a decision under constraints, use optimization.

Key Point: Most advanced planning solutions — including those built by WonForge — use both, but optimization remains the decision engine that drives measurable results.