![How does the Transmetrics forecasting product work in detail?
14
cargo demand
forecast
model
(automatic
learning and
calculation from
inputs)
Automated inputs
customer
historical data
external data
sources
(e.g. export,
import,
google trends,
holidays,
seasonality,
B2C indicators,
campaigns,
weather[?])
Manual inputs
(“adjustment levers”)
yearly orders curve
business growth %
holiday coefficients
shopping peaks
large customer growth
+ yearly curves
With limited data
(e.g. 6 months of history)
• Transmetrics suggests
initial values for the
adjustment levers from
past experience
• Customer users can
review / override values
With a big data set (e.g. 3+
years history)
• Levers ‘default setting’ is
data-mined from
historical data.
• Customer users can still
review / override them.
Other statistically
detected trends](https://image.slidesharecdn.com/6-161117102506/85/Transmetrics-14-320.jpg)
Transmetrics
- 1. The predictive optimization for cargo transport: How we can turn the Big Data into profit for business
- 2. Typical situation today - FTL • Today, due to imbalances in the flow of freight between economic regions, vehicles have to make too many empty return trips. • Across the EU, the average of empty return kilometers is 24% of total kilometers driven. • Seasonal imbalances, due to holidays, consumer behavior and unforeseen events, increase the empty kilometers even more. 2 Depot Delivery Location #2 Pickup Location #1 Delivery Location #1 Pickup Location #2
- 3. Impact in FTL and containers • 24 % empty kilometers in ground transport • Empty container flows • Low asset utilization • Trailers • Containers • Idle time of drivers away from home location 3
- 4. Typical situation today - Groupage / LTL Today, there is poor capacity utilization in the transport between terminals, for terminal – based products (Groupage or LTL), both domestic and international 4 Terminal Customer locations Terminal Customer locations
- 5. Impact in LTL and Groupage • 57% average vehicle utilization (by gross weight) • Capacity shortages around peak seasons • Uneven warehouse workloads • Impacts on customer service levels 5
- 6. Predictive analytics is critical for transport capacity optimization 6 Optimized Transport Capacity Orders for transport are received JUST IN TIME Departing trucks Have to be carefully planned in advance Visibility paradox: Planning has to be done before the data for it is available Prediction is the key to optimization
- 7. Can future transport orders be predicted? 7 0 5000 10000 15000 20000 25000 PoC data: CN-North Europe total bookings From 2013 to mid 2016 actual Major holidays Major holidays Major holidays Major holidays Season Peaks Season Peaks Season Peaks Long-term growth curve Customer Campaigns Customer Campaigns
- 8. The Big Data product of Transmetrics forecasts future shipping orders, unlocking opportunities for linehaul optimization Shipping History 3-5 years Customer orders Consolidations Linehauls Events Carrier contracts Customers ... + Shopping days Public holidays Month end Fairs and events Customer Forecasts School holidays Weekdays New tenders Network plan changes Industrial seasonality Gained and lost customers New Product launches Commo- dity prices = Efficient Linehaul plan Forecasted customer orders
- 9. Transmetrics enables the loading factor of each linehaul to be forecasted in advance (1-8 weeks before the day) Forecast: next Wednesday departure Forecast: next Thursday departure Forecast: next Friday departure Unused capacity Likely to have too much unused space: action needed Should be OK, no need for action Forecasted groupage orders via data mining Likely to be overloaded, need to make a contingency plan
- 10. Forecasting the shipment volumes enables OPTIMIZATION: tailoring the linehaul plan for each particular day Terminaldepartinglinehauls Default schedule + routing Same for every day Optimized schedule + routing, Tailored for the specific day Terminaldepartinglinehauls
- 11. Differences in shipment flows 11 Daily shipments in tons throughout the year “Winter schedule” “Summer schedule” “Winter schedule” JAN APR JUL OCT DEC Daily shipments in tons throughout the year JAN APR JUL OCT DEC “Transmetrics schedule” • Traditional network capacity planning: without prediction • Same plan every day, at most changes once per year (winter/summer schedule) • GOAL: • Adjusted plan for every day, much less excess capacity
- 12. Main levers of benefits Reduce cost for P&L impact Cancel unneeded linehauls Order variable linehauls in advance Reduce empty kilometers
- 13. Profit impact by Transmetrics Large transport company P&L structure (typical, source: Roland Berger strategy consultants) 100 Net Revenue 35 15 7 5 Local pickup & delivery Gross Profit Direct Costs Indirect Costs 3 Net Profit -3.5 +3.5 An increase in load factor of 10% can double the profit of a profitable operation, or turn around a loss-making operation 13 35 Linehaul 15 Hub
- 14. How does the Transmetrics forecasting product work in detail? 14 cargo demand forecast model (automatic learning and calculation from inputs) Automated inputs customer historical data external data sources (e.g. export, import, google trends, holidays, seasonality, B2C indicators, campaigns, weather[?]) Manual inputs (“adjustment levers”) yearly orders curve business growth % holiday coefficients shopping peaks large customer growth + yearly curves With limited data (e.g. 6 months of history) • Transmetrics suggests initial values for the adjustment levers from past experience • Customer users can review / override values With a big data set (e.g. 3+ years history) • Levers ‘default setting’ is data-mined from historical data. • Customer users can still review / override them. Other statistically detected trends
- 15. VPN Shipping history Shipments, capacities, contracts, events Transmetrics servers Transmetrics Cargo transport predictive optimization product : SaaS product with a daily usage scenario Customer IT systems Transport Management System Transport Capacity Planning System Reports for users Forecasts Optimized schedule runs periodically Cloud - SaaS Subscription 15
- 16. How is Transmetrics different from other forecasting and optimization tools? 16 Taking optimization decisions 1 month ahead of time allows for bigger changes - more P&L impact. Very granular forecasting (per depot/ZIP per day) to enable optimization. Optimization uses all detailed forecast data via a seamless integration. Optimizing 1 month ahead enables time to review, buy-in and approve changes. “Predictive optimization” – optimize based on a forecast, not on last-minute actual data. One integrated end-to-end workflow, using the same system for forecasting and optimization. Forecasting via data mining (bottom-up) improves precision at detailed levels. Utilizes external big data sources to achieve unbeatable forecast accuracy. Support strategic decisions with what-if studies, based on future - not past – data. Forecasting Optimization Predictive Optimization
- 17. Case 2: Speedy AD (DPD Bulgaria) 17 Case Study: Speedy AD (Bulgaria)
- 18. Bucharest Speedy AD overview, goals of the Transmetrics implementation 18 Sofia Vratsa Blagoev grad Veliko Tarnovo Russe Varna Plovdiv Stara Zagora Burgas Speedy AD inter-hub network: • 10 hubs • 100+ daily inter-hub truck lines • Mix of own trucks + 3rd party • 82% - 83% average load factor • Some individual departures with low load factor every day Challenges to be met with project • Support growth of shipping volumes with 3rd party capacity, without expanding own assets • Increase inter-hub load factors (to 88- 93%) • Better planning of network resources around peak seasons • Operational resilience – move the know- how from individuals to an IT system
- 19. Business processes and support by the Transmetrics product 19 Time Frame Business Process Transmetrics solution supports by: Thurdsay 1 week before 7:00 AM on the day 5:00 PM on the day Operations during the evening Morning after Plan capacity needs for next week with 3rd party transport partners Phase 2: Weekly forecast of needed inter- hub departures, based on expected shipment volumes Last minute changes to transport plan, purchase extra inter-hub trips if needed Phase 2: Daily forecast of needed inter-hub departures for express + remaining economy volume from yesterday Start consolidating and loading pickups Phase 4: Optimization of the loading and routing plan for the day, based on received shipments Execute transport plan, handle exceptions Phase 4: Changes to loading and routing plan to handle exceptions in transport Phase 1: Historical data reports on load factors, remaining shipments in transit, service levels Transport remaining economy shipments, analyze performance
- 20. Benefits of the system 20 Already achieved • Identification of historical low load factor trips, and network improvement • Identification of operational exceptions (e.g. not following loading instructions) • Continuous monitoring of network performance In progress (to be evaluated in mid - 2016) • Increase of operational loading factors from 82-83% to 88-93% • Increased network resilience, avoid under-capacity during peak seasons
- 21. Thank you for your attention! Transmetrics AD | Asparuh Koev, CEO | +359 888 400 348 | [email protected]
Editor's Notes
- #9: http://cliparts.co/cliparts/rcL/n84/rcLn84yKi.jpg