Machine learning without data science is as fast as a sports car without fuel

February 27, 2020
When you consider machine learning and algorithms to optimize your processes, make sure you build it on the right foundation.
Customer Stories
Florin Bulgarov

Machine learning. Artificial intelligence. Data science. Automation. These are so wonderful key phrases since they became the new business mantra. Machine learning is the new way for optimization, lower the costs, improve the customer experience or increase the productivity and margins. It is the tool to drive in-depth analytics and generate self improving automations. Everybody talks about it. Everybody finds it useful. Everybody wants it. And everybody says they can do it. But there’s a catch.

Machine learning learns. And to do so, it has to rely on data. Not just any kind of data, but information collected for the specific purposes, filtered through data science and made ready to be processed though computer algorithms. Otherwise, it would be just like the fastest, most beautiful, most technological endowed sports car, sitting in your parking lot with an empty fuel talk. Or even worse, a car with the engine exploded due to the wrong fuel you put in. Just like a sports car, machine learning could make you overrun anyone else, but until you give it the fuel it needs, it might become just an expensive decoration for your business.

It’s true, all businesses sit on mountains of data ready to be mined. Companies have a long history, they know their customers’ profiles, they have extensive market reports, they fine-tuned their processes in long periods of time and this expertise is gold. But, in our experience, such data usually is not aggregated, it is stored in multiple databases and application platforms which do not talk to each other. Data is not homogeneous, systemically collected and ready for analysis. In such situations, cleaning the data is a painful, long process in itself, which needs a lot of data science. Sometimes, businesses even give up and start with a clean slate, building new sets of data from scratch. Which takes us to the beginning of this article. Collecting data takes time. Learning from it takes time, too. And algorithms without data are next to useless. Unless you find available know-how specific to your needs which will help you to jump start the optimization journey.

Next to 4 Million deliveries

We, at Postis, have developed in Romania the first platform specialized in the enhancement of supply-chain processes and the last mile delivery optimization, powered by data science and computing algorithms. In the past 3 years since we commercially launched the Postis Platform, we processed nearly 4 million deliveries from warehouses to stores to end-users’ homes, with a range of products that spans from envelopes and flowers to palletized stone and industrial equipment. With data computing and statistics, big numbers really matter, while every transaction we process allows us to optimize continuously our decision making algorithms and business intelligence.

50+ options ready to deliver each and every parcel right

The Postis Platform integrates over 50 couriers and transporters able to deliver our customers’ goods worldwide, everywhere in Europe, national or at the county level. Each delivery solution has its own strengths and weaknesses, they all have different prices and geographic coverage, they provide different transport capacities, IT systems and people skills. We make sure that every time our automated algorithms process a delivery, the best solution is always selected, based on multiple selection criteria.

75+ supply-chain scenarios

We integrate in the same platform retailers, e-commerce actors, e-fulfillers, 3rd party logistics providers, tech partners, transporters and couriers and make sure that data is seamlessly handed over along our customers’ supply-chain. Together with our customers and partners, up to this moment, we developed more than 75 supply-chain optimization scenarios applicable to varied business models, industry verticals, territorial reach and levels of complexity. Such different typologies allow the Postis Platform to evolve, continuously add performance criteria and enhance our customers’ delivery engine and advanced distribution.

45+ companies rely their daily deliveries on Postis Platform

We serve big and small companies with the same level of attention. Each delivery, no matter the size, value, type of product or destination, is treated with the same care. Over 45 companies use our automated selection algorithms to choose the best transport option, and their numbers grow as we speak. Our customers are specialized in fashion, pharma&cosmetics, medical, automotive, food, home&deco, logistics, books, IT&C, electric apparel, industrial equipment and more. And, because of so many and different needs, we enable our data scientists to continuously improve our decision making machine and increase the value we provide. Starting with transport cost optimization, continuing with process automation and productivity growth, leading with business scaling, business model optimization and customer experience enhancement.

At Postis, you find the supply chain optimization engine that your business deserves. Latest technology fueled by industry specific, complex data sets, and continuously fine-tuned decision making algorithms capable to provide you the right decision, every time.
For more, drop us a mail and it will be our pleasure to assist.

portrait of man with glasses smiling

Transform your logistics and delivery through data science and state of the art digital tech.