SCHEDULE A DEMO

Driven Data Blog

Insights and information from the team at Driven Data!
 
header-graphic
Bottom-of-Chevron

Budget Optimization Engine

Posted by Jon Berna on Oct 20, 2019 9:42:22 PM

One of the most challenging aspects of marketing is establishing a budget for each channel.  Spend too little and you concede territory to the competition, spend too much and you waste resources that won’t lead to incrementally increasing sales.  The path most dealers have taken to understand this is to select an initial budget and tune accordingly each month thereafter.

Two goals:

  1. Determine the right budget to set to achieve your goals. Establish a clear cost benefit analysis showing the exact amount spending more doesn’t make sense.
  2. Output a prediction on every car in your inventory to proactive boost slow turning vehicles.

There is a problem though…

Each month thereafter your dealership’s inventory, process and strategy changes.  So has the competition and their strategy.  The initial budget you set is now massively outdated and numb to everything that’s changed.

Driven Data’s Budget Optimization Engine an industry 1st solution. 

Budget Optimization Engine

We approached solving these problems from two perspectives.

  1. How to choose the right budget?
  2. How can I act on this information?

To determine the correct budget, we first had to understand the influence ad spend has on sales.  Leveraging our universal architecture, we measured the impact ad spend has on inventory turn.  We chose turn, versus sales because it more accurately produces the result the dealership wants.  Sales output is just a function of how fast each individual vehicle sells in the aggregate.

Every dealer’s data is different, as are the cars they stock and sell, so our engine accounts for these variations using dynamic input feature selection.  This is just a smart way to avoid the noise and instead focus on the signal in every dealership’s data that makes them unique.

Brand Level Prediction

BOE Brand

Model Level Prediction

BOE Model Level

By combining the ad spend data with the variables from the inventory we build a prediction of how fast each vehicle will sell given different budgets.  We use 4 different models and select the best one based on its accuracy.  You can now see at various levels of spend on average how fast your inventory will turn.

But how else can you act on this information?

The old way of inventory targeting was to increase ad spend on cars as they age.  The problem with this approach is that the car is already old.  By waiting until the car is old you focused on the symptom not the disease.

With Driven Data we automatically increase ad spend on cars that ‘are predicted to age’ which allows you to avoid them from getting aged in the first place.  We now can create strategies to target your vehicles based on how fast we predict they will sell.

Every night Driven Data refreshes the prediction.  It takes each variable into consideration i.e. (price, VDPs, leads, cost) and predicts a new expected date of sale.  Now all the VIN based ads are automatically adjusted to spend the optimal amount for that day.

This allows our dealer partners to choose the correct budget and make the most of it every day.

Topics: Marketing, Product, Machine Learning, Paid Social, Paid Search

Join the dealerships selling and servicing more cars with Driven Data.

SCHEDULE A DEMO