The Automated Quotes use case (use case is equivalent to a project) is now eight months old and we want to celebrate it with a place in the spotlight. The use case was initiated for several reasons: 1) The original quoting process is a manual and time-consuming process, 2) it is conducted differently across the business and 3) data is not collected and used to optimise the process.
As a team that loves data, the last point really hurts, which was why it was decided that a change was needed. Together with Logistics, we wanted to reduce the time-to-market for quotes, we wanted to standardise the quoting process across the entire Logistics business, while making sure that everything we do is as transparent as possible. And of course, we wanted to collect, track and measure every quote that we produce.
Enter: Automated Quotes
The idea was simple: Can we use historic data to produce automated quotes? (If you haven’t guessed it, this was where we got the name for the use case). So, what is an automated quote? Based on very little input from the user (usually a sales person), we can, by utilising AI (machine learning), estimate the different cost elements (i.e. loaded haulage cost, empty haulage cost, crossing cost) that makes up the total cost of a booking from a collection area to a delivery area. Combine this with a pre-set margin and a final sanity check from the sales person and you have yourself a very effective cocktail: An automated quote.
We started out on the Vlaardingen and Immingham corridor, where we worked closely with Perry Schalker and Sarah Holloway to make sure that we didn’t create something with no value for the business.
Did we succeed? (Hint: Yes)
Automated Quotes is now being used every day in Vlaardingen and Immingham where, on average, it is being used by eight different people creating around 50 quotes per day and around 4,100 quotes since July 2018. It has fully replaced the old way of quoting for full-loads.
Hamburg has just completed a testing period and will go live in January. We continue to scale and will shortly decide on the next office in line. Furthermore, the solution has proved to be more accurate in predicting the cost for a quote than the previous quotation tool. For example, instead of calculating the expected empty haulage for a booking as a national average, the solution automatically finds the closest 300 historical bookings and uses their actual empty haulage to estimate the empty haulage for a given quotation.
Unfortunately, without performance data from the time before Automated Quotes, it has not been possible to create a proper benchmark analysis. What we can do is to share a story from real life telling one of the changes the tool has brought:
• Previously, the good sales people in Immingham were not allowed to quote from the Netherlands to the UK without receiving confirmation from the Vlaardingen office. It went like this:

• Now, because the quoting process has been standardised and the quote amount is the same independent of who creates it, this is the process:

Sarah Holloway, UK Commercial Director, says: “The tool is fantastic! For the first time in a long time it is easy to make a quote, it takes just seconds to do and the customer has it in their inbox fast meaning our chances of success are higher at winning more business for DFDS. We now use it for tenders also which has significantly reduced our time spent so we can be more focused on the solution rather than figuring out the price.”
The next important step for us is to start utilising the data we are currently collecting for every quote made. We can soon analyse whether we always lose quotes from one region or whether it is the same customer who is always asking for quotes but never gives us any business.
Thank you to Rasmus Fisker Pedersen, Project Manager in Smart Data, for sending this very interesting story.