Predicting and mitigating the out-of-stock problem for retailers.
Every year in the $4.5 trillion US retail industry, retailers lose a tremendous amount of revenue to Out-of-stock (OOS). It’s a problem that’s plagued the industry for decades, but because the solution has always been viewed as more expensive than the problem itself, it’s gone unchecked.
NextOrbit uses advanced data sciences and a number of machine learning techniques to reduce OOS and drive higher revenues. When a customer knows what they want, they can rest assured that they’ll find it at any retailer that utilizes the NextOrbit platform.
And retailers can finally breathe a sigh of relief.
This happens every day, in every store across the world. And every time it happens, that retailer loses an average of 40 percent of what Jane would’ve spent. Manufacturers lose another 35 percent.
How often does it happen? According to industry research, retailers experience an OOS rate of eight percent.
That eight percent costs retailers a loss of eight to ten percent in sales. For promoted items, the numbers are even higher: as high as 15 percent.
And the losses don’t stop there. Industry research also shows that the third time a customer is faced with an out-of-stock item, they’ll go elsewhere. So retailers aren’t just losing on the first missed transaction: they’re potentially losing out on every future transaction from that customer.
OOS impacts every facet of the retail experience:
Retailers have been aware of the OOS problem for decades, and they’ve been content to simply chalk it up as a loss and call it a day.
Why? Because the solution—predictive analytics—is seen as more expensive than the problem. If that’s the case, why would they fix it? It’s more cost-effective to simply let disappointed shoppers walk out the door.
So the question remains unanswered: what could we have sold if we could solve the out-of-stock problem?
Distribution centers make it their priority to stock well; it’s what they do, so they typically invest in smart systems that help them stay well-stocked. So if distribution centers are well-stocked, why can’t stores keep their shelves stocked accordingly?
What happens between the distribution center and the retail store that’s causing such disastrous losses?
Most retailers were forced to upgrade their platform when the Y2K bug was an imminent threat. Fast forward nearly 20 years later, and most are still using that same platform.
It’s no wonder that those platforms are useless when it comes to accurately predicting their stocking needs- they precede the iPhone by a full eight years.
Most Computer Assisted Ordering (CAO) systems used by retailers to place orders from their distribution centers rely on history data to keep their shelves stocked accordingly.
But CAO only takes into account a small subset of relevant data: they don’t have access to other factors like local events or weather, let alone the crucial data provided by predictive analytics.
The other solution relied on by retail stores is the electronic shelf sensor, which identifies OOS as it happens; when a store runs out of an item, smart shelves take note.
But smart shelves can cost upwards of $200,000 for a 14,000-square-foot installation, and it doesn’t help stock shelves. It only notifies the store of the problem. It doesn’t solve it.
NextOrbit predicts when an item will go out-of-stock.
It also tells retailers why an item is out of stock, so retailers can take action before their customers look elsewhere.
NextOrbit’s consumer-facing app will tap into the vast database created by the NextOrbit platform to alert shoppers to store availability and even help them plan their shopping trips with the confidence of knowing that if NextOrbit says it’s in stock, it’s in stock.
NextOrbit uses predictive analytics and data sciences like Time Series Algorithms, Kalman Filtering, and machine learning to create a top-down overview of exactly what’s happening with each and every purchase:
- Point of Sale data
- Store receipt
- DSD receipt
- Store orders
- Inventory Position
- Historical error trail
- TPR, Markdowns
- Weather data
- Macroeconomic data
- Local holidays and events
- Social feed and blog
- Competitors’ promotion strategy
All that data means NextOrbit can take action as soon as it’s necessary. With our streaming POS feed, if a store hasn’t sold any milk for two hours, the staff can immediately be notified to address the problem.
Fine Grain Detail
Using more input than ever before possible, NextOrbit can provide predictions on an individual SKU level. The results are impressive.
All that money that’s been walking out the door for so long can now remain in retailers’ stores. When everything runs more efficiently, customers find the products they want, retailers keep the customers they have and the $4.5 trillion retail industry suddenly runs like a well-oiled machine.
- CIOReview’s “Top 20 Companies Enhancing Customer Experience”
- Retail Business Technology Expo’s “Top 16 Most Innovative Companies in UK”
- “RetailCIOOutlook’s “Top 10 Retail Cloud Platforms”
NextOrbit is a complete, fully functional product that’s ready for market. We already have five client conversions pending in the US, three in the EU and one in India. We’ve signed a strategic alliance partnership with Win Weber and Associates to take us to market in the US, and scaling will start by the end of the year.
Ready to hear what’s next for NextOrbit? Click the “business profile” section at the top of this page to find out more about the NextOrbit opportunity!
Kishore Rajgopal, Founder & CEO
Kishore has more than 20 years of experience in consulting, product development, business development, business intelligence and analytics. During his nine years at Infosys, he developed technology consulting and technical architecture that repositioned Infosys as a consulting player. He also served as Global Head of Business Intelligence & Analytics at HCL and co-founded Crowdanalytix.
Mayank Shende, Cofounder and Data Scientist
Mayank has patented a data-driven methodology to determine quantitative and qualitative trade-off between risk and profitability. He’s a recognized expert in Process Optimization, Pattern Recognition, Kalman Filtering, Dynamic Programming, Model Predictive Control, Computational Techniques, Artificial Neural Networks and Frequency Analysis.