+15 %
versus the benchmark
when comparing forecast quality against a major planning system.
When standard planning logic is no longer enough
Forecasting alone does not make planning reliable. We close the gap between forecast and decision-making.
We develop tailored forecasting and AI solutions for complex supply-chain decisions and turn reliable forecasts into planning logic that works in practice alongside SAP IBP, OMP, Kinaxis, and existing ERP and planning systems.
scroll to read more
For companies with complex planning processes, long lead times, broad portfolios, and demanding service-level targets, we build reliable specialized logic exactly where standard systems already exist, but have to deal with real data, real constraints, and real trade-offs.
From forecast to reliable planning.
In genuinely complex planning environments, we do more than add custom logic beyond the standard. We make sure forecasts actually translate into reliable planning decisions.
The results our customers achieved.
This is what our customers have achieved with our support.
+20 %
planning improvement
compared with the company’s previous planning process.
View success storypatent application
resulting from a forecasting project.
What could be possible in your environment? Our calculator highlights a typical economic lever in 30 seconds and makes the potential of complementary planning logic tangible.
Determine your potential nowYour standard planning works. Just not for every case.
Modern planning systems are strong when processes are stable, enough history is available, and standard logic applies.
Challenges tend to show up in two places: first, in genuine special cases that go beyond the standard, such as new products without history, intermittent spare-parts demand, long replenishment times, volatile markets, or unusual influencing factors. Second, where domain-specific fine-tuning determines planning quality: forecast level, hierarchies, method selection, special effects, manual overrides, and service-level logic.
With demo data, a large planning solution can look convincing very quickly. Real data is where you find out whether the logic truly holds up under actual conditions. That is when the real work begins.
Standard planning
- Stable core processes
- Items with long history / NOS items
- Standardized forecasting logic
- Robust rule-based replenishment
Demanding planning cases
Cold Start
Slow Movers
Hierarchies
Special effects
Excel workarounds
Manual overrides
Special logic
Indicators
Long lead times
Intermittent demand
Long tail
prognostica as complementary planning logic
We complement planning systems exactly where standard logic reaches its limits and where domain-specific fine-tuning determines whether planning becomes truly reliable.
From forecast to planning
Good forecasts are only the beginning. What matters is turning them into planning decisions.
The real value often does not come from forecasting alone, but from the question of how a forecast becomes effective inside a planning system. That is where much of the domain-specific and methodological customization actually sits.
That is why we take a more holistic view of planning strategy. In the end, what matters is not just forecast accuracy, but whether the logic delivers the desired results under real conditions, for example in service levels, inventories, or planning effort.
We do not start from a blank slate: we know the typical pain points of S&OP practice, from manual overrides and coordination loops to trade-offs between availability, inventory, and operational feasibility, and we address them directly in the planning logic.
Choose the planning level
Item, customer, region, or hierarchy: the right level determines forecast quality, stability, and usability.
Set methods and parameters appropriately
Intermittent time series, long-tail items, or special effects require different logic from clean standard segments. It is equally important to decide how forecasts should be evaluated in the first place and which accuracy metrics fit the planning case.
Evaluate influencing factors correctly
Promotions, market shifts, manual overrides, and external indicators need to be incorporated into planning in a way that makes business sense.
Safeguard planning impact
A planner navigation system segments planning objects deliberately: where is manual adjustment useful, and where can the planner rely on automated planning? Only then do forecasts turn into reliable decisions.
Specialized extensions for SAP, OMP, or Kinaxis.
We step in where standard approaches reach their limits: in forecasting questions that require deep mathematical, statistical, and process understanding.
Our solutions intentionally start as a lean layer alongside your existing system landscape. That lets us build a reliable prototype with real data quickly, without touching your core system. Once the value is proven, the solution can be expanded step by step, run productively, and integrated into your ERP or planning systems where needed.
Fast to validate
Prototype with real data instead of months of concept work.
Mathematically grounded
Custom models instead of standard parameters.
Open for integration
Start as a stand-alone solution, then expand later through interfaces.
Implementation with future
With future, move faster from logic to implementation.
Not every demanding planning case has to be built technically from scratch. Depending on the use case, we either develop the required forecasting and decision logic individually or use our software future as an accelerator. That gets us from the planning question to a reliable prototype with real data much faster.
future helps us deploy proven forecasting building blocks, data logic, and integration patterns exactly where they make your planning process more robust. At the same time, it leaves enough room for the special logic your standard system does not model properly today, whether for spare parts, cold-start situations, long-tail items, hierarchies, or unusual influencing factors.
For you, that means less technical rebuild, faster validation, a clean path into productive use, and, if needed, gradual integration into your existing system landscape.
Learn more about future as an implementation foundation
Faster to a reliable prototype
With real data, proven building blocks, and a clear focus on the concrete planning case.
Less technical rebuilding
Existing components from future shorten the path from idea to usable solution.
Cleanly integrable
Start as a lean layer alongside the standard system, then expand and connect it through interfaces as needed.
Typical demanding planning cases
Four situations where complementary planning logic creates real value
01
New products without history
Cold Start
New products
No history
Spare parts
We use product structures, bills of materials, metadata, semantic similarities, and historical comparison items to generate reliable initial forecasts before classical forecasting methods even have enough data to work with.
Typical question: Which spare parts for a new product should we stock initially, and in what quantity?
02
Intermittent demand and long-tail items
Long Tail
Slow Movers
Intermittent demand
Inventory risk
Many spare parts, C-items, or service items have rare and irregular demand patterns. We model demand probabilities, consumption patterns, uncertainty, and service-level effects to enable better decisions on inventory and availability.
Typical question: How do we secure availability without overstocking the long tail?
03
Special effects, events, and manual planning logic
Special effects
Promotions
Indicators
Marketing spend
Manual overrides
Special logic
Promotions, product changes, market shifts, customer-specific exceptional demand, or one-off effects distort historical data. Planning systems such as SAP allow values for marketing effects or similar drivers to be entered, but in practice it is often unclear how these values should be chosen in a meaningful way. With our support, or through an additional planning logic layer, this effect can be determined on a mathematically sound basis, if needed using market, industry, and marketing indicators.
Typical question: Which effects belong in the forecast, which should deliberately be excluded, and which additional indicators help evaluate promotions and market changes realistically?
04
Forecast level and hierarchical forecasting logic
Hierarchies
Aggregation
Forecast Level
Reconciliation
SAP
Before a planning system can produce a good forecast, it often first needs to be clarified at which level forecasting should happen at all: item, product group, region, customer, or another hierarchy level. We analyze demand behavior, data quality, and aggregation logic to determine the right forecasting level in a reliable way and, if needed, also handle reconciliation across hierarchy levels.
Typical question: At which grouping level should we build the forecast so that both the planning system and the planner receive truly reliable results?
Example calculation for a typical ROI lever in the specialty segment
Make inventory risk and tied-up capital visible in 30 seconds.
In just a few seconds, assess how much capital you currently tie up in specialty segments to protect against uncertainty, long replenishment times, and high service requirements.
If your forecast error in this segment improves by , you will immediately see how much safety capital could potentially be released at your selected service level.
This calculator deliberately shows only one exemplary ROI path through safety capital. Additional effects such as time savings, fewer manual interventions, and better decisions in demanding cases are not included in this calculation.
Mathematical logic
Safety stock = z × forecast uncertainty × √lead time
Conservative ROI assumption
Selected service level, assumed error reduction, and valuation of released capital based on WACC.
Capital currently tied up in the specialty segment.
Delivery performance you are currently protecting in this segment.
Approximation for MAPE/WAPE in your specialty segment.
Average replenishment time in months.
WACC as the annual financing effect on released capital.
Expected improvement in forecast error through the specialized planning logic.
Live result
Your current forecast ties up capital to buffer uncertainty. Assuming an error reduction of , the following potential emerges:
Potentially releasable capital
This corresponds to of the segment inventory under consideration.
Annual cash-flow effect
Interest or financing effect based on your WACC.
Optimized forecast error
Error reduction through a specialized forecasting layer.
Safety capital tied up today
Safety capital tied up with prognostica
Next step
Request the detailed report to get a clearer view of the full potential. Then let us use the follow-up use-case check to assess how robust this logic is for your specific planning case and which data and process levers would have the strongest effect.
The calculator deliberately shows only one slice of the picture. In practice, the actual value of an individual planning logic often comes not only from lower capital lock-up, but also from less planning effort, fewer mistakes in exceptional situations, and better decisions under uncertainty. That is exactly why the next step focuses on your concrete use case, your data, and the actual business impact.
Book a use-case checkExample use case
Use case: Forecasting new spare parts before product launch
A typical application is planning spare parts for new appliances with no historical demand, long lead times, and high service-level pressure.
New devices create new spare parts. Classical forecasting methods can hardly produce reliable values when there is no history. At the same time, parts often need to be ordered early because lead times can be several months.
A separate planning logic analyzes the bills of materials of new appliances, finds semantically and structurally similar parts in existing devices, and derives failure rates and initial forecasts from them.
Planners receive not just a number, but a transparent recommendation: comparable proxy part, matching rationale, historical consumption basis, derived rate, and recommended initial quantity.
Three benefits
Cold start solved
Forecasts are created before any demand history exists.
Planners stay in control
Every recommendation is explainable and can be adjusted manually.
Scalable instead of Excel workarounds
BOMs, metadata, consumption histories, and appliance forecasts are connected systematically.
Your challenge is probably different.
That is exactly why we do not begin with a standard solution, but with a structured use-case check. The case shown above is an example of what custom planning logic can look like. The real task is always to understand the logic required for your specific planning case, test it with your data, and make it visible in a reliable prototype.
From the use-case check to a productive custom planning solution.
Our approach is intentionally pragmatic: first narrow down your question properly, then build a prototype with real data, and finally scale and extend it based on planner feedback. We do not even have to touch your core system for that. Only later do we integrate the solution into your existing planning system. That turns an idea into a reliable decision process instead of a major transformation project.
1
Use-case check
Identify the most urgent pain point that fails in the standard setup.
Request a use-case check2
Pilot / MVP
Create a solution for selected key users or a defined subset of the data.
3
Scaling & stabilization
Extend to all relevant data, tune the models, and improve UX and processes.
4
Integration
Seamless connection to SAP IBP, OMP, or Kinaxis.
AI provides the evidence. Humans make the decision.
Especially in critical planning decisions, a black-box forecast is not enough. Our solutions are intentionally designed to keep responsibility with the planner.
We place high value on explainability: every recommendation is shown together with its reasoning, data basis, and comparison logic. Planners can review, adjust, and comment on suggestions. These adjustments then feed back into performance tracking.
The result is a system that does not just generate forecasts, but continuously learns where AI performs well and where human domain knowledge makes the difference.
FAQ
Can’t SAP do this as well?
SAP IBP offers powerful forecasting functions, machine learning, and options for integrating external algorithms. Our approach does not contradict that. It makes deliberate use of exactly this openness.
What matters, however, is this: a powerful planning system alone does not guarantee reliable planning. In practice, the key challenge is often the domain-specific fine-tuning of the logic, for example at the forecast level, in hierarchies, for intermittent demand, special effects, or particular planner workflows.
prognostica develops the complementary planning logic for cases that cannot be solved cleanly with standard parameters or generic forecasting models, such as BOM-based cold-start forecasting, custom failure-rate models, semantic parts matching, or specialized decision and planning logic.
We are not competing with classic SAP customizing. We complement existing planning systems exactly where methodological depth, real data complexity, and economically relevant special cases become decisive.
We intentionally start pragmatically alongside the existing system. Once value is proven, the solution can be integrated into the existing SAP, ERP, or planning landscape.
When does complementary planning logic pay off?
Complementary planning logic pays off above all when a clearly defined planning case currently causes a lot of manual effort, keeps coming up in discussion, or has a noticeable impact on inventory, availability, or decision quality.
Typical cases are those in which standard systems generally work well, but individual constellations still have to be handled through Excel, rules, estimates, or planner experience. That is exactly where the value of additional planning logic can usually be made visible very quickly.
What data do we need to get started?
That depends on your case. For an initial prototype, existing operational data is often already enough, such as sales or consumption history, material master data, bills of materials, product and hierarchy attributes, inventory information, or planning values.
What matters is not perfect data completeness at the start, but whether the relevant decision logic can be represented and meaningfully tested with real data. That is exactly what we assess in the use-case check.
How much integration effort is required?
At the beginning, deliberately very little. We typically start as a stand-alone solution alongside the existing system landscape so that business teams and planners can validate quickly with real data, without triggering a large IT project right away.
Once the value is proven, the solution can be connected step by step, for example via file imports, APIs, BI pipelines, or direct handovers into SAP- and ERP-adjacent processes. The integration architecture follows the business case, not the other way around.
Does the planner stay in the loop?
Yes. That is especially important in critical planning decisions. Our solutions are designed to support planners with evidence, comparison logic, and transparent recommendations, not to hide decisions inside a black box.
The planner can review, adjust, and contextualize suggestions. The result is a system that is both more mathematically grounded and better connected to real operational planning.