Forecast Module - WeYield Apps

Demand forecast

Have better understanding of your future performance to be more confident of your pricing and fleet decisions.

The Forecasting AI accurately predicts (average 94%) future booking and utilization situation based on current and past levels of performance.

Understand your future bookings and utilization levels for the next 30 days (more options in the Paid version).

Understand the accuracy of future predictions and how the forecasted levels have evolved over time.

Create your own forecast filters such as your top region/station, car types, car groups, segments, and brands (paid option).

Select your own days in advance to forecast. For example, default is 3, 7, 14, 21, 28 days in advance of the pickup date.

To understand more the philosophy followed, watch Christian Cadéré presentation made during our last London Academy and learn more about this new module.

Forecasting advantages

Artificial Intelligence-based forecasting tends to significantly more accurate than the current ‘Estimation’ methodology. 

  • More precise than the estimate, taking account of the current sales trends or exceptional events.   
  • Forecast uses full data history, seasonality and recent changes in booking pace, whereas the estimate only uses the past single year’s performance.
  • Machine learning methodology is constantly learning and refining, thus improving accuracy over time.

How it works

  • A.I. based methodology looks for correlations between past on rent values while considering the following parameters: Year, Month, Day of Week, Weekend or midweek, Holidays/events, Date.
  • A.I. model considers all historical data (trends/seasonality/noise) and tests several machine learning methods and chooses the best for each client.
  • The forecast is produced for the next 30 picks up days (or customerdefined period).
  • WeYield displays the forecasted on rent values from 5 advance pickup dates (default is 3, 7, 14, 21, 28 days or customer defined date intervals) and their past accuracy for clients to review performance.

Included with piloting

  •    Global view
  •    5 dates intervals
  •    Algorithm choice and fine tune
  •    Forecast history reconstruction
  •    Geography, brand, cars or customer filters

Paid option

  •    Global view
  •    5 dates intervals
  •    Algorithm choice and fine tune
  •    Forecast history reconstruction
  •    Geography, brand, cars or customer filters

Frequently Asked Questions

Can I get the forecast module activated if I only have rateshoping?
No, you need your demand and fleet module activated thus an existing data feed The report is activated in my account but there are only a few data points? Once activated the report will start generating data points each night and this will take time to build up sufficiently to provide forecasting lines.

Are you sure that the new report will provide better accuracy than the existing estimate feature?
Great question, the A.I algorithms will learn from your own data and will improve in accuracy over time. The estimation is a rough calculation based on your current position this year and the same position last year taking into account your final position. While the estimate can be accurate most of the time, the new methodology takes into account many other data points to forecast your final position, for example, we find the estimate performs very well during ‘flat’ periods. When there is an activity shift due to changes in key dates (eg Easter is on a different weekend) or a significant trend change compared to last year, then the forecast performs better as it takes this variance into account.

Can you forecast more than 30 days out? Can I see station level forecasts? Or Car group level forecasts?
Yes for all, however, please note that the further the advance forecast is, the less accurate the results become. Both paid and free versions allow a greater than 30 day forecast period. The paid version of the report allows clients to determine the filters for the nightly forecast. 

Why do we not get an accuracy % figure for future forecasts?
Because an accuracy ratio can only be computed when we get actual results. We only show accuracy figures for past data as we compare the forecast to the final figure achieved (actual). The future forecast shows a % of utilization, according to the fleet plan.

logo la french tech

We have the French Touch!
WeYield is 100% developed, financed,
supported and hosted in France.

...

Due to ITB cancelation, our Revenue Management Forum will become an online webinar instead. More information to come on our social networks and by email. Stay tuned.