Online returns have been a challenge for retailers since the beginning of eCommerce. This is because of both their volume compared to normal stores, and the costs associated with processing them. In the wake of COVID-19, this problem could prove more critical then ever, as online becomes retailers’ singular sales channel. There is already early evidence of this, with preliminary data from Quantum Metric showing that eCommerce associated with Brick and Mortar retailers saw an average revenue weekly growth rate increase of 52%, and Nike Inc.’s digital sales went up by 36%.
The eCommerce returns dilemma
Online shopping is popular for a reason, but the convenience and choice of eCommerce comes at the price of not being able to ‘try before you buy’ for customers.
This simple difference is the reason online returns are so much more prevalent than for Brick and Mortar stores. In eCommerce the customers’ homes becomes the fitting room. And, just like any fitting room, products end up back on the shelves. According to Happy Returns, while shoppers return only 10% of what they buy in stores, they send back up to 50% of what they buy online.
This is compounded by customers accounting for this when ordering online. A survey from Barclays found that 30% of shoppers deliberately over-purchase and subsequently return unwanted items. Additionally, 20% regularly order multiple versions (often sizes) of the same item so they could make their mind up when they are delivered, all of which is facilitated by the retailer at great cost.
While it might seem logical to have stricter returns policies, or make customers cover the cost of returns, consumer expectations make this a risky strategy. According to the 2017 UPS Pulse of the Online Shopper survey, 68% of shoppers view returns policies before making a purchase. This leaves retailers with a catch-22 situation when it comes to losing out on online sales or losing profits from processing the inevitable returns that comes with those sales.
What are the challenges of retail returns?
So why are returns such a strain on retailers?
Cost of returns – First and foremost is the simple cost of returns. Since returns are in themselves essentially lost sales, the added cost of returning them, which according to CNBC is on average 30% of the purchase price, can heavily impact retailer’s margins.
Processing returns and reverse logistics – On top of this is the resources and effort of processing returns and getting the stock back available to be sold as quickly as possible. This reverse logistics can be particularly challenging and can result in returned stock not being available for purchase again for some time, often leading to out-of-stocks on the webshop. According to the Barclays report, 57% of retailers say that dealing with returns has a negative impact on the day-to-day running of their business.
Contamination concerns with COVID-19? – A unique and recent challenge, particularly for apparel retailers, is dealing with the potential contamination and contact of returned good with the COVID-19 virus. Initial research suggests that the virus can only survive on fabric surfaces for 24 hours, but for up to 72 on plastics like packaging. This will need to be addressed by eCommerce retailers who continue trading throughout the epidemic.
Return fraud – This is a challenge shared by brick-and-mortar stores. Fraudulent returns cost the US alone 27 billion dollars a year. This can involve the ‘returning’ of stolen merchandise for cash, stealing receipts to enable a false return or using someone else’s receipt to return unpurchased store stock. Naturally, using receipts for returns presents a risk, APPRISS found that receipted returns are more than twice as likely to be fraudulent as other methods.
How to reduce the impact of returns on eCommerce
So what options are there for retailers looking to tackle their returns problems?
- Reduce the likelihood of returns, without harming customer experience or sales: Include accurate and detailed product descriptions. Use uniformed/standard sizes where possible and provide a more specific sizing filter. Offer virtual ‘try-ons’ with augmented reality/3D imaging.
- Set clear and accessible rules regarding returns: Make sure customers know what and how they are allowed to return items, this reduces spending resources processing illegitimate returns.
- Improve visibility: Maintain a single view of stock with item-level inbound and outbound processes, this will also allow for online returns back to stores, and ship-from-store. Make this visibility accessible to your entire team and your customers.
- Improve efficiency of inbound and outbound processes: Utilise reliable & efficient technologies and automated processes like exception handling. One of the leading technologies for this is RFID, which prevents the need to open any boxes or packages as it can count and verify items without direct line of sight.
- Improve internal processes: Ensure returns processes (and supporting software) enables additional layers of merchandise management such as grading items based on quality and tracking when the item was returned.
- Counter Return Fraud: Verify legitimacy of returns as much as possible, best practise involves unique item-level validation like RFID or unique serial numbers.
- Ensure returned stock is safe to sell: Implement processes to ensure returns are not damaged in any way, implement a policy to account for safe handling of merchandise during theCovid-19 pandemic, either sanitising products or leaving them a set amount of time before adding back into webshop stock.
Using RFID tagging and looking to improve return processes?
NEW: eCommerce returns module
eCommerce/ DTC is increasing due to COVID-19. This causes an over proportional increase in returns which normally would require a ramp up of staff and equipment to handle the process – Detego’s new RFID enabled return process provides an approx 90% productivity increase.
Want the latest retail and retail tech insights directly to your inbox?
Visual Merchandising’s Data Problem
Visual merchandising is one of the ‘dark arts’ of retail. For the uninitiated, it’s the practice of designing visually appealing sales floors and store fixtures that attract customers and ultimately sell the products on display. It’s a subjective and artistic means of delivering a concrete KPI – sales.
One of the challenges of visual merchandising is data, experimentation & design is all well and good, but not if you have no way of knowing what does and doesn’t work. Sales data is always a good place to start, but there are so many other factors at play that it’s often unwise to attribute visual merchandising to an increase or decrease of sales to a single product. You have to go one step further…
What is Money Mapping?
So, how do you measure the sales performance of a store based on its layout and design?
First, you map out the sales floor with the exact location of products. Then you break the map of the store into ‘zones’, typically around certain fixtures, shelves and displays. These zones can consist of several different items, grouped by either category or style depending on the design of the store.
You then collate the sales of every item in this zone and compare it to others in the same store. The result? An impression of shop floor sales broken down by areas of the store.
This is called ‘Money Mapping’ and allows retailers to visualise and analyse which areas of a store are ‘hotspots’ and which are ‘cold’ in terms of sales. This gives an initial view of which areas and fixtures are selling products and which aren’t.
To account for other external influences on sales, best practise is to swap items between fixtures or observe a ‘money map’ over a long time, as collections and merchandise changes between seasons. This way, if the localised sales data remains relatively similar even after products have rotated, then it’s clear the design or locations of the fixture is having an impact.
What are the benefits of Money Mapping?
- Insight on consumer experience
- Provides valuable data for visual merchandisers
- Breaks down areas of sales floor by sales performance
- Can be used to optimise store layout
- Drives Sales
- Can compare Product Placement & Visual Merchandising
- Can be used to conduct A/B tests
This all sounds great, so why doesn’t every retailer and every store do this already? The simple answer – the process of matching the sales data to specific locations on the sales floor, manually for every item and every store, is logistically a big ask. This means, if this can be done at all by visual merchandisers, it can only be done in a small number of stores.
How does AI change money mapping?
So how can we solve this data problem for visual merchandisers and make ‘Money Mapping’ easier and more accessible for retailers?
The first issue is having an accurate map of a store which includes exactly where every single item is sold from. Traditionally this would have to be done manually, and then have the sales data of items cross-referenced with their location in a store.
The solution: Using RFID (Radio Frequency Identification) and AI localisation techniques, we can now create a map of a store as part of the daily or weekly stock count.
This is done by adding ‘reference’ RFID tags into the store. Small tags just like ones that go on products are placed on fixtures and walls in the store. Because these never move, we can use the signal strength (relative to the fixtures) from stock counts to map exactly where items are in the store and what items they are grouped with.
This location info is then integrated with data from point of sale to generate an automated Money Map of a store, as part of the regular reporting and analytics function of the store. This can be done for as many stores as desired. With the data collection automated, visual merchandisers can focus on using the data to optimise product placement and store design across stores.
With larger data sets to work with, this also opens up the potential for more detailed analysis and experimentation such as A/B testing product combinations and store layouts!
What’s the process for AI Money Mapping?
- Attach reference RFID tags to walls and fixtures within the store
- Perform regular RFID stock takes as normal
- Software uses machine learning to ‘map’ out item locations within the store
- Integrate point of sale data with RFID software
- Software produces ‘heat map’ of the store based on sales
- Visual merchandisers can use data to inform strategy and measure results
Visual merchandising is a subtle but valuable process for retailers. Done properly it has a huge impact on both sales, customer experience and brand image. The only problem with this is visual merchandisers often don’t have enough data to measure performance and identify where their attention is needed most. The data they can collect is either time consuming, expensive or inaccurate.
Artificial intelligence changes the game for visual merchandising. By utilising RFID tags and Machine Learning, it is possible to ‘map out’ the location of items in a store, and more importantly, the sales distribution of the shop floor. These ‘Money Maps’ tell visual merchandisers what areas of the sales floor are ‘hotspots’ for sales and which are underperforming. Using this data, they can then focus their attention on improving the design or layout of certain areas.
Additionally, with this data stores can look to leverage their sales hotspots either by prioritising the best locations for best-sellers, high-value items or items that are due to go out of season.
Either way, AI-enabled Money Mapping is another important evolution in retail data and analytics. Providing retailers with unprecedented insight into exactly what goes on in Brick and Mortar Stores.
Want the latest retail and retail tech insights directly to your inbox?
When it comes to using Artificial intelligence (AI) and Radio Frequency Identification (RFID) in retail for process optimisation, the majority of use cases involve management or ‘HQ’ level decision making. These include automating functions such as store planograms and stock optimisation between stores.
However, AI can also impact retail on a much more micro and everyday level, actively assisting store staff in one of their most common daily routines –stock replenishment.
Using RFID and the information it collects from stock reads, we can produce AI pick lists to optimise and guide staff through the replenishment process. Not only are we combining RFID technology and AI algorithms to produce these pick lists, but existing RFID processes are already assisting staff. When you put all this together, replenishment become a walk in the park.
Let’s start at the beginning…
RFID-based Stocktake and Replenishment – The backbone of modern stock management
With RFID, store staff can do regular (often daily) cycle counts of the entire store quickly and easily. This is simply done by walking around the backroom and salesfloor with a handheld reader that counts items that are several feet away, using radio frequency. An RFID application or software, like the Detego platform, will then compare the actual stock levels of the shop floor with the desired counts (i.e., planogram), and tell staff exactly what needs to be replenished from the backroom.
So far, what has just been described has been entirely RFID-based and is the standard process for RFID in retail. This is already far easier and more accurate than traditional methods, not to mention the actual effect of the technology like higher stock accuracy and product availability. But why stop there?
Taking it one step further – AI pick lists for ‘mapping’ the perfect replenishment path
Normally, even with the support of RFID, the store staff are then left to fulfil replenishment by themselves, using the list provided by the application. These pick lists are often only sorted by product features such as name or price. Because back rooms can be quite large in bigger stores, or densely packed in smaller ones, the staff’s ‘pick path’ can be incredibly sporadic. This is made even worse in the case of new staff who don’t know the layout of the backroom by heart, or even experienced staff if stock has simply been moved around and updated with the start of a new season.
By utilising new tag localisation techniques, it is now possible to locate where items are in the backroom in relation to each other. This is done during the regular RFID stocktakes that are already taking place, utilising data mining and machine learning pipelines without any need for additional hardware or specialist tags. Using this information, we can create automated AI pick-paths that, using a mobile application, guide staff through replenishment and present the most efficient order to collect items in.
The above example is designed to present the quickest possible replenishment route for staff, so is solely using items’ distance from one another to calculate a pick list. However, AI pick lists can process the replenishment list in a number of ways, depending on what the store wants to focus on.
Replenishment paths could take additional factors such as product value or expiry date into account, alongside the location of the items. It would then look for items that fit this rule and are nearby one another in the backroom. For example, a pick list targeting on-floor-availability would group nearby items that are running low on the sales floor, so that these items are refilled first to speed up the replenishment process whilst also combatting loss of sales from out of stocks.
Benefits of AI-pick lists
Artificial intelligence (AI) is becoming increasingly utilised within the retail industry. One of the main challenges with the technology is having enough relevant data to be utilised or ‘fed into’ an AI system. With Radio Frequency Identification (RFID) providing huge amounts of accurate item-level data for merchandise, the two are a match made in heaven.
This webinar cover AI’s applications for stock optimisation and how machine-learning can ensure products are always in the right place at the right time, including:
- AI-driven, automated planograms for optimised product availability
- Visual merchandising and ‘Money mapping’ in stores to monitor and increase sales.
- How RFID stock-takes and AI pick-lists combine to make replenishment faster and more accurate than ever
- How machine learning smart fitting rooms are bringing accurate cross-selling into the physical store
- Using AI for demand prediction and stock optimisation across store networks
Digitally native competitors and demanding customers are forcing a new perspective in retail. Artificial intelligence and machine learning have huge implications for technology and is one of the main driving forces of the ‘fourth industrial revolution’. The AI in retail conference explores how this fast-emerging technology is changing the retail landscape.
The event takes place on the 16th of October at Cavendish conference center in London (W1G 9DT).
With speakers from several major retailers and brands including Sainsbury’s, Domino’s, Microsoft and google, as well as our customers Levi’s, the conference promises to be incredibly insightful on the practical applications of AI in retail.
Presentation: AI for inventory and stock optimisation – ensuring the right products are in the right place at the right time
Our senior data scientist, Simon Walk, will be discussing how the Detego platform offers AI capabilities to retailers as part of a SaaS solution. The presentation will cover AI’s applications for stock optimisation and how machine-learning can ensure products are always in the right place at the right time, including:
- How Artificial Intelligence works for stock optimisation
- AI-optimised planograms – How AI and machine learning can be utilised to produce more up-to-date and relevant planograms that are optimised for each individual store
- Machine-Learning Product recommendations – How AI is used to provide superior product recommendations to customers to increase sales, and how Detego utilises this technology in smart displays to bring cross-selling to brick-and-mortar stores
- Smart Picking Lists – How AI can take stock replenishment to the next level by guiding staff through replenishment with optimised ‘picking paths’ that calculate the most efficient replenishment route for store staff
Who should attend the AI in retail conference?
The event is aimed at senior executives within retail in the following departments:
- Digital Transformation
- Data Science / Data Engineering / Data Analytics
- Sales & Marketing
- Customer Services
Coming to the conference and want to schedule a meeting? Get in touch below:
The retail industry has been subject to enormous change in recent years, a trend that looks likely to continue even now retail has found its feet in the digital age. This significant shift in the landscape coupled with the collapse of numerous household brands naturally created something of a panic in the industry and a fear an impending ‘retail apocalypse’. So, has this taken place and are brick and mortar stores heading toward extinction in the age of online shopping? The short answer: no. Retail sales, in general, are increasing, and whilst online sales are certainly growing at a faster rate than in stores, in the UK market Brick and Mortar stores make up roughly 84% of all sales, a trend that applies globally.
In fact, with the help of new technology, in particular, AI brick and mortar stores can take lessons from e-commerce and bring new innovations into the physical store. A study by Capgemini found that around 68% of pure-play online retailers have implemented artificial intelligence in some fashion, compared to only 10% of brick and mortar stores. Utilising new technology can often be the key to growth, so perhaps it’s no surprise that online retailing is growing at such a faster rate. This is beginning to change however, as physical stores are learning to adapt.
AI-assisted cross-selling in the store
The ability to cross-sell using AI is a huge strength of online retail. Online stores use machine learning to make intelligent and tailored product recommendations to customers while they are shopping This naturally increases sales, but does so relative to the quality of the product recommendations, hence the use of AI to perfect the process and make more successful recommendations. Brick and mortar stores have traditionally never been able to make use of cross-selling like this, having to rely on in-person customer service to drive sales. But as technology has improved, they can now make use of both.
The main reason for this is the emergence of chatbots as a mobile platform for in-store product recommendations. Whilst chatbots themselves originate from online, certain retailers now implement them in their stores to assist customers. Their primary function is often to assist customers if store associates are all occupied. Chatbots can help customers locate items, find more information on products and check stock availability, all through their smartphones. Intelligent AI-powered chatbots will then be able to recommend products similar to the item’s customers were searching for, not only increasing sales but improving the customer experience.
The fashion industry has taken this concept even further with the use of smart mirrors. Smart mirrors use advanced tagging technology (RFID) to sense the items that a customer brings into a fitting room. It will then display the products on the mirror and, like a chatbot, assist customers by providing information on other available sizes and colours and recommending products that are often bought with the ones being tried on. Most smart mirrors also have a feature that calls store staff to assist the customer, either providing face-to-face support or bringing the recommended items the customer has chosen to the fitting room. Smart fitting rooms are the epitome of the merger of the digital and physical in brick and mortar stores.
AI-powered in-store business intelligence
Artificial intelligence can also go a long way to revolutionising traditional brick and mortar store processes. One of the best examples of this is the use of AI assisted planograms. A planogram is essentially a plan of where items should be displayed on a shop floor to maximise customer purchases. In certain sectors such as fashion, planograms must be even more detailed and include the optimum display quantities of each different size and colour.
Artificial intelligence can revolutionise the planogram, using machine learning to constantly optimise not only the positioning of merchandise but the most efficient quantities of different articles to display on the shop floor. The advantages of this process being performed through AI are huge. Not only does it largely automate the process, saving time for staff, but it will constantly adapt and improve and can personalise and optimise the planogram on an individual store level.
The future of retail, online and in-store
To conclude, the impact of digital technology on the retail industry, in particular brick and mortar stores, has been significant but not actively catastrophic as many feared it would be. The emergence of ecommerce has gone from a threat to a strength for some physical stores. Not only has omnichannel retailing (something we did not have time to explore, but we go into detail on here) allied the different methods of shopping, but brick and mortar stores have now begun to incorporate certain technologies from online.
AI technology, originally something only online retailers could really utilise, is now finding its way into brick and mortar stores and improving both store processes and customer experience. We are also just scratching the surface of AI’s uses in retail and as more retailers choose to adopt the technology more benefits will be discovered. So, in reality brick and mortar stores are far from going extinct; they are in fact evolving and will continue to do so.
Over the course of recent years, Radio Frequency Identification (RFID) technology has been applied in many different business domains to handle inventory challenges, for example, the problem of monitoring stock. Retailers can use the knowledge of item locations to understand in which area of the store they can locate articles, such as the backroom or sales floor. By possessing this knowledge, retailers can rely on having optimal product availability for their customers through an efficient replenishment process. When using Detego’s InStore software, RFID technology provides the benefit of up to 99% stock accuracy and on-floor availability. However, there have always been limitations with the ability to understand an item’s exact location, due to the long-distance read signal an RFID tag emits. The most common “best-practice” methods that can ensure such high figures are currently known to involve a physical method of shielding between individual stock locations, whether this is using foil or metallic paint. However, this common and costly method of shielding has left retailers and our data scientists wondering what else could be done to lower the hurdles on the initial steps of the RFID journey.
Beating the laws of Physics
Detego’s data scientists have debated and solved this common problem by utilising machine learning and tag localization to establish the ‘Smart Shield’ feature to determine the relative locations of individual items. When RFID was developed, the concept of identifying individual tag locations was never considered a possibility. However, the innovation and progress made by our team has already defied the laws of physics, with our ‘Smart Shield’ results showing performance to be on par with physical shielding requirements. By removing this barrier, a much faster and affordable implementation into brick-and-mortar stores has been made possible.
How it works
Every individual RFID tag creates a specific time & signal stamp when it is read by an RFID reader – up to several times per second. Using the data that this process collects, we apply smart algorithms to our software so that it can continuously learn and improve the accuracy each time a stocktake is performed without physical shielding. Traditionally, ‘fixed reader infrastructure’ was the only technique of identifying individual item locations without physical shielding. To combat the rigidity of this method, we modified this approach to handheld readers to generate a highly cost-effective and scalable solution.
What it means
No more physical shielding – Firstly, we can solve one of the major downfalls of RFID by eliminating the necessity for physical shielding between a retailer’s separate stock locations. In saying that, ‘Smart Shield’ is not the silver bullet that 100% removes the need for physical shielding. There will be challenging stores where Smart Shield runs into limitations. However, there are also stores where it’s difficult to apply physical shielding – so here the ‘Smart Shield’ can help.
Continuous improvement with Artificial Intelligence – Applying our machine learning algorithms to the data will result in automatic learning & improvements with on-going stocktakes; further improving accuracy & efficiency of store processes.
More than just ‘shielding’ – This new capability for RFID tag localization has huge benefits far beyond the removal of physical shielding in retail stores. Through the machine learning infrastructure in place that enables us to identify locations of individual RFID tags, we can improve operations in the following ways:
- Replenishment advice and picking lists for e-com processes can be optimised through location-based sorting to make the process more efficient for the store associate when in the backroom.
- Staff will now additionally be supported in finding items to further enhance customer service.
Utilising the information about relative locations of items can drastically increase staff efficiency and business process efficiency across various functions relating to logistics, warehousing and retail. Not only will this save their budget, but they will also experience faster time to market and their franchise/wholesaler partners will be enabled in a more streamlined process that eases the implementation of the new technology.
The use cases of RFID technology have progressed even further than initially anticipated. From the first use case to our Smart Shield feature using artificial intelligence. The latter is now readily available as part of our platform and retailers can easily utilise the new system, seeing an increase in process efficiency across both the backroom and the shop floor. But perhaps the biggest development enabled by the ‘Smart Shield’ is the greater ease with which retailers can now implement RFID systems in stores. As this relatively-young technology continues to improve, developments like the ‘Smart shield’ are further steps towards RFID becoming part of the status quo in the fashion retail landscape, as advancements such as this discover new use cases for the technology whilst simultaneously reducing the cost and time required for implementation.
“Utilizing the information about relative locations of items can drastically increase staff efficiency and business process efficiency across various functions relating to logistics, warehousing and retail. Together with features that will significantly lower the costs for a roll-out, retailers will experience an even faster time to market and will have the ability to enable franchise/wholesaler partners with full RFID capabilities at low costs.”
Retail software specialists, Detego, have presented their ground-breaking methodology for in-store product recommendations, helping bring the same quality of cross-selling over from e-commerce and into the physical store. The new AI-based recommendation engine will enable retailers to provide personalized product suggestions utilizing data unique to store locations and point of sale information, without the need for identifying customer profiles.
Cross-selling through related product recommendations has always been a huge strength of e-commerce, with 35% of Amazon’s revenue generated by its recommendation engine (source). In recent years, innovations in RFID-based solutions such as smart fitting rooms and mobile chatbots have opened the doors to automated product recommendations within physical stores. Whilst the technology is now available, there is still one more hurdle between Brick and Mortar stores and effective cross-selling. This is namely the fact that the best recommender systems require vast amounts of both personal and aggregated data to provide effective suggestions, and whilst this is at a surplus in e-commerce, physical stores traditionally struggle with data being limited as well as sparse.
Speaking at the ACM UMAP 2019 in Cyprus in June, data scientists from Detego, who specialise in RFID-based software solutions for retailers, presented their proposed method of data-manipulation for in-store recommender systems with a paper titled: ‘Beggars Can’t Be Choosers: Augmenting Sparse Data for Embedding-Based Product Recommendations in Retail Stores’. The approach involves an alternative algorithm that leverages shopping-baskets and common-item combinations combined with point of sale information. Detego says this allows retailers to provide targeted recommendations with a 6.9% increase in quality, aimed at individual stores, without having to maintain separate models for each location. When combined with the technology to deliver these product recommendations, retailers could see a substantial increase in sales in Brick and Mortar stores, whilst customers will see a more connected and engaging in-store experience, as Detego continues to bridge the gap between online and the physical store.
“Customers who bought this also bought…” is no longer a phrase reserved exclusively for customers of e-commerce platforms. Due to the adoption of RFID-based technologies, such as Detego’s Smart Fitting Room, personalised recommendations can also be presented to customers of brick and mortar stores. Moreover, Detego’s AI-based recommendation engine is tailored towards the specific requirements of fashion retail stores, such as fast-changing and varying product assortments.’ says Matthias Wölbitsch, Detego data scientist.
With Detego now successfully rolling out the Smart Fitting Room application alongside their real-time inventory management software, this latest improvement is another opportunity for retailers to evolve their stores for the future.
Planograms are a key part of running a retail store. In simple terms, they dictate the what, where and, the how many of products on the shop floor.
This article explores planograms, the challenges they present and how retails ongoing digital revolution along with artificial intelligence is changing planograms as we know them.
What is a planogram?
A planogram defines the location and quantity of products to be placed on display. They essentially function as blueprints for all merchandise within a store.
Alongside the visual merchandising element, planograms aim to optimise article availability and thus specifically stimulate sales.
So, a store planogram has two aspects:
1 – The Merchandising: which articles are presented on the sales floor and how (visual merchandising)
2 – The detailed quantities for individual colour and sizes (availability optimisation)
For example, a planogram might state where a formal shirt is displayed on the shop floor. The planogram will also state what quantities should be displayed, specific to different sizes and colours. So the planogram may state 3 mediums, 2 smalls, 2 larges and 1 extra-large are available on the shop floor, and so on.
Ideally, merchandising and planograms go hand in hand: customers are inspired by the presentation and then, their desired product is available in the matching size. The reality, however, often paints a different picture.
Planogram – Merchandising
Planogram – Quantities and Priorities
Visually appealing and available in relevant sizes
What are the challenges of setting and maintaining a retail planogram?
Breaking this down, there are two questions that retailers face:
1) How do you define a planogram for my stores with a suitable size distribution?
The first question can be a challenge for retail depending on your definition of suitable. The reality is simply that manually creating and defining different size distributions for every single product is incredibly time-consuming and rarely worth the effort. This means generally stores will have a single set size distribution or ratio for every product in the store. This certainly could be worse, but it could (and can) be better….
2) How do you maintain the planogram on the shop floor? I.e. how to ensure that the products on the planogram are always on the shop floor
This, on the other hand, is a major challenge for retailers. What good is a planogram if it’s not adhered to? For example, if a planogram states that a size XL should be on the shop floor at all times but it is missing due to inaccurate or slow replenishment/refill, and a customer who needs that size can’t find it, more often than not that sale is lost.
There are two ways steps that go a long way to fixing this:
- Use a more specific and data-driven planogram in the first place (What we’re going to explore now)
- Having a better replenishment process in the store. (Like this one)
How does AI change the planogram?
Manual maintenance and adjustment of specific planograms is rarely a realistic option for retailers, especially when it comes to specific size distribution for individual products and stores.
By utilising artificial intelligence and machine learning procedures, it is possible to automate this process to define a precisely optimised size distribution for all articles across the store.
Not only does this save an enormous amount of planning time, but it also addresses the ongoing dynamics in individual stores. The self-learning system adapts to possibly changing conditions and continuously optimises the plan.
- Produces automated planograms based on sales data
- Produces optimised planograms for individual products
- Produces optimised planograms specific to the individual store
- Constantly adapts and optimised the planogram based on new data
AI planograms in action with the Detego platform
If this all sounds theoretical, its not.
During the operational process in the store, Detego InStore also supports the store personnel at several occasions: The software offers two parameters that provide information about article availability at any time and therefore represent important KPIs:
- On-floor availability: The percentage of all available articles that are currently displayed on the sales floor
- Planogram compliance: Provides information on how well the planogram with its individual size distribution is implemented on the sales floor
If one of the two parameters fall below certain threshold values, store staff needs to action: In addition to classic ERP systems, Detego InStore offers a finer level of granularity in the stores, by telling store staff that certain articles are available in the backroom but not on the salesfloor and therefore need to be refilled to comply with the predefined planogram.
Retailers benefit from a complete process for the planning and implementation. Another advantage: Refill advices in the app are sorted such that the search in the back room is made as efficient as possible by minimising walking routes.
With AI planograms, shelf space is used for top sellers and is not wasted on sizes that are rarely or never bought. With its self-learning components, the Detego platform for the store makes a suitable proposal for all sizes and facilitates implementation in daily processes – including relevant KPIs for measuring performance. And if a certain size is not available in one store, the platform offers an exact inventory view of surrounding stores – ready for click & collect.
Benefits for retailers:
- Individually optimised AI planogram per store
- Efficient use of shelf space according to bestsellers per store
- Guided processes: from planning to refilling
- KPIs to provide insights on operational excellence per store – in real-time
Benefits for consumers:
- High on-floor availability for the locally popular sizes
- Positive customer journey
- Overall increased article availability through exact inventory data on the entire store network – including reservation options
There’s a lot of talk about artificial intelligence at the moment, but not that many practical use cases, especially in bricks-and-mortar retailing. We should know. We were one of the early adopters of machine learning in the development of our retail software and have launched a popular chatbot tool to help consumers get quick answers to simple stock enquiries without having to seek out a sales assistant or call customer service.
Aside from a general foreboding that there’s more talk about AI than action, and that machines might eventually outperform humans, we get the distinct impression that most retailers are still trying to understand the basics about what machine learning can actually do for their business and how AI will help with consumer engagement. So, here’s our attempt to unravel the mysteries behind the topic of artificial intelligence in retail and see whether it’s a passing fad, or something retailers should be paying more attention to.
Knowing your product
There is one field where artificial intelligence has already made a significant difference, and that’s getting to grips with vast amounts of data and making much more informed product recommendations. It’s a technique that was spearheaded online by the likes of Amazon – making suggestions about what other products you might like – but still has a long way to go on the high street. There just isn’t a very joined-up approach between the worlds of online and bricks-and-mortar retailing. And the availability of products is often neglected; not many retailers take into consideration what products would be best (or most profitable) to shift.
That’s where AI comes in.
Artificial intelligence, by amalgamating lots of data and making decisions based on a variety of factors – product availability, purchase history, current trends, profitability, and so on – gets better and better at making reasoned choices. All this might sound obvious, but it’s something very few retailers actually do in their stores. Most retailers still have rather disjointed processes across various channels and different departments suffer from a lack of data input on a consistent level. Some processes are semi-automated, but many – such as merchandise planning and product assortments – remain largely manual. For example, buyers still base most of their decisions on out-of-date sales figures and gut instinct, rather than using much more efficient machine learning tools.
Artificial intelligence is ideally suited to forecasting and stock allocation. These processes historically tend to be quite manual and cumbersome and generally are not managed that efficiently – largely because it’s just too much work to find the ideal mix. It’s something that’s typically done by relatively small departments, even though product selection and stock availability are clearly so fundamental to a retailer’s bottom line. Yet self-learning mechanisms can be put in place to maximise availability and promote what’s most likely to sell.
In most stores today, store staff are prone to stack shelves based on whatever available sizes there are and what will fit, rather than having technology that knows what will be best for that particular store’s profit. By using AI, we’ve found that different stores, even in the same town, might require completely different sizes of garments to maximise their sales.
Coupled with RFID tags on every item to help monitor stock with near hundred percent accuracy, artificial intelligence can be used to improve on delivery performance and the distribution of inventory between stores, the warehouse and even to the consumer. For example, we found one retailer flummoxed by hundreds of cartons of products having been shipped but never received and various departments spending a long time trying to resolve – something a machine could unravel in seconds. It’s not uncommon for retailers to not know exactly where stock is at any particular time, but the joy of artificial intelligence is ensuring that human mistakes are minimised and promises – such as a marketing campaign offering the latest product release at a special price in chosen stores – are always fulfilled.
While there’s still a lot of hype about chatbots at the moment – robots to help with customer service – and undue pressure to adopt some of the latest (and arguably more advanced) techniques inherited from the online retail world, this is just the start of what can be done regarding artificial intelligence in retail.
Chatbots, available on people’s smartphones for asking advice when they enter a store, can give an instant link to customer services and access to a much wider catalogue of products and services. The appeal of chatbots has not only come about because young people, in particular, are now used to dealing with them online; but also because a lot of people don’t really want to have to track down a sales assistant in a busy store and then get the ‘don’t know’ or ‘don’t care’ response that can ruin a retailer’s reputation. Simple, product-related tasks and stock checks are ideally suited to machines. We’ve even found sales conversion rates to be up to five percent higher in stores where we’ve introduced chatbots with AI-powered product recommendation tools.
Clearly, chatbots won’t ever replace humans working in stores. But they can certainly complement them: often proving more reliable; never prone to sickness or breaks; not trying to drag you into a conversation and oversell, and always available 24/7.
Increased stock transparency and profitability, as well as better customer engagement, are what will help keep retailers flourishing – allowing them to compete with the increasing threat of online retail goliaths who are setting the trends for the future of shopping. Artificial intelligence isn’t just another fad. It might seem like fiction, but it’s clearly here to stay, outperforming expectations all the time.