by Valentyn Kropov, Associate VP of Client Success at SoftServe
It is obvious that retail is a highly competitive environment. Not only retailers compete with each other redrawing the market as they usually do. The bigger and more influential battle reveals between online and offline. E-commerce gained 10% market share in the US in 2018. The figure is not that big, but taking into account its rapid growth (13.7% estimated till 2021) and ambitions, it does give a hard time to brick & mortar retail, whose margin is low enough to feel the consequences of such a solid customers’ outflow. So, there is no way to stay still in such an aggressive environment. Thus, brick & mortar retailers, such as Walmart and Target, launch online capabilities with quick delivery and click & pick options available. At the same time, a unique advantage of brick & mortar lies in giving customers the options to test or try out prior to purchase. Thus, Best Buy, one of the largest electronics retailers in the US, joins efforts with the biggest producers (Apple, Samsung, etc.), who split the rental fee, to be able to maintain its stores and sustain competition with Amazon.
As to the deployment of innovative solutions, retailers tend to be careful. They observe and examine best practices rather than embrace every emerging novelty. But it’s just impossible to keep up to date and enhance business performance without it.
Innovations drive progress and boost major changes in all the realms, and retail is not an exception.
Robotics to transform the way retail operates
There is a chain of complicated processes that precede the final action – a consumer takes a product from a shelf or receives an online order. Now this process in general looks like this: production – warehouse – order fulfillment – delivery. Yet, it can definitely be transformed and shortened. Today most of the processes involve human labor, which is time and resource consuming. Its automation will save labor costs, improve productivity and consumer experience.
As of now, production is the most automated stage and there is a reason for that. No matter how complicated an item is – a car, a cellphone or a pen – its production requires an exact set of actions to be done under the same circumstances. In other words, the level of uncertainty is pretty low: you know everything about the product, its size, materials, environment where you build it – it helps a lot to automate production processes. These are the most favorable conditions for automation. Subsequent stages are more challenging because uncertainty is growing during warehouse, order fulfillment and delivery stages.
The next step of automation that we are currently working on, will concern warehouse storage and fulfillment. The unknown are the size and features of the goods that robots deal with. For example, one order can contain a vase and book. How can robohands identify that a vase has to be transported carefully in order not to be broken and that the book pages can be ripped? And there are hundreds of thousands of possible combinations. This is the issue, that is being addressed now. Amazon’s warehouses are equipped with over 100, 000 robots, that are used to carry stock around the expansive warehouse floors and group together all the individual items needed for a specific order. It reduces but doesn’t exclude, the number of interactions humans have with the products. So, along with machines, 500, 000 employees maintain Amazon warehouses. Startups are more flexible in terms of innovations roll out, as it’s easier to set the new system than update an existing one. That’s why smaller and newer companies like Ocado beat the behemoth and created a fully automated human labor-free warehouse.
The last but not least task is to deliver a packing to a consumer. While international transportation is relatively cheap and quick, last-mile delivery is a real tall order. It takes 80% of the whole delivery cost to have a courier get an order from a warehouse and bring it to a consumer. Automation of this part and replacing human by a machine comes down to unpredictable obstacles faced on the way from point A to point B – weather conditions (sun, haze, rain, snow etc.), type of pavement, ladder etc. It affects and sometimes even blocks, robots’ operation. For instance, in the summer of 2018 we launched smart parking at SoftServe Headquarters in Lviv that proved to be extremely effective at finding free parking spots and saving drivers’ time. While the weather was nice and shiny, it worked smoothly. But once it started snowing, we had to retrain computer vision algorithm so it could operate under different weather conditions. Last-mile delivery faces the very same challenge. A lot of conditions have to be foreseen. But a solution is possible to find. Amazon has already announced drone delivery of purchases. We’ll see, where it will get them and what will happen when it will start snowing in Seattle.
Automation of production, warehouse management, order fulfillment and last-mile delivery itself is not a destination point for automation. It’s more of a start for the next era of retail.
The future retail model
Not all of the elements of the recent production & delivery chain, are necessary. After full automation some elements will be extra, so the next step is optimization. For example, there is no need to store big amounts of items in warehouses. Ordered goods can be delivered to our homes right from a manufactory. It will be possible firstly because of acceleration of the whole process and secondly due to open demand prediction possibility, which will allow any retailer to calculate how many items will be purchased within any time span and plan production or procurement of this very amount beforehand.
The next step is to eliminate long-distance deliveries. In the future manufactories will be built around key metropolises. Now it seems surreal, but 3D printing technology will evolve into a huge industry, first dealing with plastic and then with metal.
So, this new chain will consist of 2 elements only, instead of 4: production and delivery. Once we order some, for instance, tableware, it will be produced on the nearest plant and delivered straight to our home by a drone.
And the final step of retail transformation is creating a cloud, similar to a digital one, to conduct a wide range of operations accompanying production and sales. Whereas now the developing of a scheme, negotiating with a contractor, who often resides in another country, and making payments take an enormous amount of time and effort, in the future, it will be done by few clicks in the cloud. Such a retail model will be the shortest, the quickest, and, consequently, the cheapest.
Different customer experience
No matter what innovative solutions appear or how the retail model evolves, one thing remains constant – the client is at the core of the business. So, along with the business processes optimization, there is another essential direction, where innovations come in handy and make a real difference – customer experience.
In the course of time customers become more and more demanding to shops. Now it’s not enough to offer a wide range of high-quality products. The retailers are expected to foresee and exceed clients’ needs. And technologies are key to success.
SoftServe’s approach is addressing the market’s most topical issues in the most sophisticated, logical and effective way. Even if some solution already exists, but we see a more comprehensive way to go, we are willing to do it. One of the explicit examples is personalization, which has become a vital instrument in retail. Currently, it operates in quite a plain way – consumers are grouped in accordance with purchase history. The algorithm assumes that if some of group members ordered some item, the rest may be also interested in this very purchase. This approach is sometimes called collaborative filtration. It’s not always efficient because true personalization is possible when you understand persona, not only what he or she clicks on. So SoftServe team went further and developed algorithm, that is able to get holistic psychometric profile of consumer based on Big Five/OCEAN traits: Openness, Conscientiousness, Extraversion, Agreeableness, Neuroticism. The data is gathered from different sources: social media accounts, shopping habits, purchase history, etc. Having processed data and created a person’s profile, the tool suggests the type of advertisement that is likely to work the best for this very person. Another solution enables a quick accurate visual search among the variety of products of similar types. Moreover, we plan to launch another demand prediction solution before long.
All of such tools will make shopping easier, less stressful, and time-consuming for the people and help the retailers adjust product range and shop merchandising to the target audience demand.
Overall, the future of retail is all about personalization, automation, and more accurate demand estimation. Machine learning will identify our needs and preferences once we get to a shop or enter website. That is definitely a win-win strategy for both – a seller and a consumer.