As per yesterday’s post, any supply chain planning solution developed before 2010 isn’t necessarily built on a modern multi-tenant cloud-ready SaaS stack (as such a stack didn’t exist, and it would have had to be partially to fully re-platformed to be modern multi-tenant cloud-ready SaaS). Any solution built after was much more likely to be built on a modern multi-tenant cloud-ready SaaS stack. Not guaranteed, but more likely.

KetteQ‘s Adaptive Supply Chain Planning Solution is one of these solutions that was built in the modern age on a fully modern multi-tenant cloud-native SaaS stack, and one that has some advantages you won’t find in most of the competition. I was able to get an early view of the latest product which was released last week. Founded in 2018, ketteQ was built from the ground up to embody all of the lessons learned from the founders’ 100+ successful supply chain planning solution implementations across industries and systems, and the wisdom gained from building two prior supply chain companies, with the goal of addressing all of the issues they encountered with previous generation solutions. The modern architecture was purpose built to fully utilize the transformational power of optimization and machine learning. It was a tall feat, and while still a work in progress (as they admit they currently only have three mature core modules on par with their peers in depth and breadth [although all inherit the advantages of their modern stack and solver architecture]), but one they have pulled off as they can also address a number of other areas with their other, newer modules, and integration to third party systems (particularly for order management, production scheduling, and transportation management) and address End-to-End (E2E) supply chain planning, with native Integrated Business Planning (IBP) across demand, inventory, and supply — which are their core modules, along with a module for Service Parts Planning and S&OP Planning.

In addition to this solid IBP core, they also have capabilities across cost & price management, asset management, fulfillment & allocation, work order management, and service parts delivery. And all of this can be accessed and controlled through a central control tower.

And most importantly, the entire solution is cloud native, designed to leverage horizontal scalability and connectivity, and built for scale. The solution is enabled by a single data model that can be stored in an easily accessible open SQL database, in a contemporary architecture that supports all solutions. The solution is extendable to support scalability, multiple models, multiple scenarios per model, and a new, highly scalable solver that can perform thousands of heuristic tests and apply a genetic algorithm with machine learning to find a solution by testing all demand ranges against all supply options to find a solution that minimizes cost / maximizes margin against potential demand changes and fill rates.

Of course, the ketteQ platform comes with a whole repertoire of applied Optimization/ML/Genetic/Heuristic models for
demand planning, inventory planning, and supply planning, as well as S&OP. In addition, because of its extensible architecture, instead of manually running single scenarios at a time, it can run up tothousands of scenarios for multiple models simultaneously, and present the results that best meet the goal or the best trade-off between multiple goals.

KetteQ does all of this in a platform that is, compared to older generation solutions:

• fast(er) to deploy — the engine was built for configuration, their scalable data model and data architecture make it easy to transform and integrate data, and they can customize the UX quickly as well
• easy to use — every screen is configured precisely to efficiently support the task at hand, and the UX can be deployed standalone or as a Salesforce front end
• cost-effective — since the platform was built from the ground up to be a true multi-tenant solution using a centralized, extensible, data architecture, each instance can spin off multiple models, which can spin off multiple scenarios, each of which only requires the additional processing requirement for that scenario instance and only the data required by that scenario; and as more computing power is required, it supports automatic horizontal scaling in the cloud.
• better performing — since it can run more scenarios in more models using modern multi-pass algorithms that combine traditional machine learning with genetic algorithms and multi-pass heuristics that go broad and deep at the same time to find solutions that can withstand perturbations while maximizing the defined goals using whatever weighting the customer desires (cost, delivery time, carbon footprint, etc.)
• more insightful — the package includes a full suite of analytics built on Python that are easily configured, extended, and integrated with AI engines (including Gen-AI if you so desire), which allows data scientists to add their own favorite forecasting, optimization, analytics, and AI algorithms; in addition, it can easily be configured to run and display best-fit forecasts at any level of hierarchy and automatically pull in and correlate external indicators as well
• more automated — the platform can be configured to automatically run through thousands of scenarios up and down the demand, supply, and inventory forecasts on demand as data changes, so the platform always has the best recommendation on the most recent data; these scenarios can include multiple sourcing, logistics, and even bills of material; and they can be consolidated meta-scenarios for end-to-end integrated S&OP across demand, supply, and inventory
• seamless Salesforce integration — takes you from customer demand all the way down to supply chain availability; seamless collaboration workflow with Salesforce forecast, pipeline, and order objects in the Salesforce front end
• AWS nativity — for full leverage of horizontal scalability and serverless computing, multi-tenant optimization and analytics, and single-tenant customer data. Moreover, the solution is also available on the AWS marketplace.

In this coverage, we are going to primarily focus on demand and supply (planning) as that is the most relevant from a sourcing perspective. Both of these heavily depend on the platform’s forecasting ability. So we’ll start there.

Forecasting

In the ketteQ platform, forecasts, which power demand and supply planning,

• can be by day, week, month, or other time period of interest
• can be global, regional, local, at any level of the (geo) hierarchy you want
• can be category, product line, and individual product
• can be business unit, customer, channel
• can be computed using sales data/forecasts, finance data, marketing data/forecasts, baselines, and consensus
• can use a plethora of models (including, but not limited to Arima[Multivariate], Average, Croston, DES, ExtraTrees, Lasso[variants], etc.), as well as user defined models in Python
• can be configured to select the best fit algorithm automatically based on historical data, based on just POS data, POS data augmented with economic indicators, external data (where insufficient POS data), etc.

These models, like all models in the platform, can be set up using a very flexible and responsive hierarchy approach, with each model automatically pulling in the model above it and then altering it as necessary (simply by modifying constraints, goals, data [sources], etc.). In the creation of models, restore points can be defined at any level before new data or new scenarios are run so the analyst can backtrack at any time.

Demand Planning

The demand planning module in ketteQ can compute demand plans that take into account:

• market intelligence input to refine the forecast (which can include thousands of indicators across 196 countries from Trading Economics as well as your own data feeds) (and which can include, or not, correlation factors for correlation analysis)
• demand sensing across business units, channels, customers, and any other data sources that are available to be integrated into the platform
• priorities across channels, customers, divisions, and departments
• multiple “what if” scenarios (simultaneously), as defined by the user
• consensus demand forecasts across multiple forecasts and accepted what-ifs

The module can then display demand (plans) in units or value across actuals, sales forecasts, finance forecasts, marketing forecasts, baseline(s), and consensus.

In addition to this demand planning capability and all of the standard capabilities you would expect from a demand planning solution, the platform also allows you to:

• Prioritize demand for planning and fulfillment
• Track demand plan metrics
• Consolidate market demand plans
• Handle NPI & transition planning
• Define user-specific workflows

Supply Planning

The reciprocal of the demand planning module, the supply planning module in ketteQ leverages what they call the PolymatiQ solver. (See their latest whitepaper at this link.)

Their capabilities for product and material planning includes the ability to:

• compute plans by the day, week, month, or any other time frame of interest
• do so globally, regionally, locally, or at any level of the hierarchy you want
• and do so for all regional, local, or any other subset of suppliers of interest, as well as view by customer-focused dimensions such as channel, business unit and customer
• use the current demand forecast, modifications, and taking into account current and projected supply availability, safety stock, inventory levels, forecasted consumption rates, expected defect rates, rotatable pools, and current supplier commitments, among other variables
• run scenarios that optimize for cost and service
• coordinate raw and pack material requirements for each facility
• support collaboration with suppliers and manufacturing
• manage sourcing options and alternates (source/routes) for make, buy, repair and transfers

Moreover, supply plans, like demand plans, can be plotted over time based on any factors or factor pair of interest, such as supply by time frame, sourcing cost vs fill rate, etc.

In addition, the supply planning module for distribution requirements can:

• develop daily deployment plans
• develop time-phased fulfillment and allocation plans
• manage exceptions and risks
• conduct what-if scenario analysis
• execute short-term plans
• track obsolescence and perform aging analysis/tracking

Inventory Planning

We did not see or review the inventory planning module in depth, even though it is one of their three core modules, so all we can tell you is that it has most of the standard functionality one would expect, and given the founder’s heritage in the service parts planning world, you know it can handle complex multi-echelon / multi-item planning. Capabilities include:

• manage raw, pack and finished goods inventory
• set and manage dynamic safety stock, EOQ, ROP levels and policies
• ensure inventory balance and execution and support for ASL (authorized stocking list), time-phased, and trigger planning by segment
• support parametric optimization for cost and service balancing
• the ability to minimize supply chain losses through better inventory management
• the ability to optimize service levels relative to goals

Salesforce: IBP

As we noted, the ketteQ platform supports native Salesforce integration, and you can do full IBP through the custom front-end built in Salesforce CRM, which allows you to seamlessly jump back and forth between your CRM and SCM, following the funnel from customer order to factory supply and back again.

The Salesforce front-end, which is very extensive, supports the typical seven-step IBP process:

1. Demand Plan
2. Demand Review
3. Supply Plan
4. Pre IBP Review
5. Executive IBP Review
6. Operational Plan
7. Finalization

… and allows it to be done easily in Salesforce design style, with walk-through tab-based processes and sub-tabs to go from summary to detail to related information. Moreover, the UI can be configured to only include relevant widgets, etc.

In addition, users can easily select an IBP Cycle; drill into orders and track order status; define custom alerts; subscribe to plans, updates, and related reports; follow sales processes including the identification and tracking of opportunities; jump into their purchase orders (on the supply side); track assets; manage programs; and access control tower functionality.

As a result of the integration with Salesforce objects, including Pipeline and Orders, the solution helps bridge the gap between sales and supply chain organizations, enabling executive-driven process change. As an advanced supply chain solution on the Salesforce Appexchange, it enables the broad base of Salesforce customers on the manufacturing cloud a slew of unique integration possibilities.
And, of course, if you don’t have Salesforce, you still have all this functionality (and more) in the ketteQ front-end.

Finally, the platform can do much more as it also has modules, as we noted, for service parts planning, service parts delivery, sales and operations planning, cost and price management, fulfillment & allocation, asset management, clinical demand management, and a control tower. It is a fundamentally modern approach to planning that is worth exploring for companies that are challenged to adapt in today’s disruptive supply chain environment. For a deeper dive into these modules and capabilities, check out their website or reach out to them for a demo. This is a recommendation for ANY mid-sized or larger manufacturing (related) organization looking for a truly modern supply chain planning solution.