Home » Here
Shelfcheck (beta), on-shelf availability optimization
A
webapp that monitors the on-shelf availability of your products at store level and that delivers
prioritized alerts when products go out-of-shelf. Full historical analysis of past OOS events is also delivered, offering both
insights and benchmarking capabilities.
Best suited for retailers. No statistical skills required. Suitable for both extremely large and small companies. Shelfcheck imports your POS (Point-of-Sales) data directly from your existing business apps and generates web reports. Get the alerts on your smartphone too.
Shelfcheck differs significantly from
Salescast.
Shelfcheck reacts to inventory problems that cannot be identified by simply looking at electronic stock records, either because records are inaccurate, or because people can't find products at their proper locations. In contrast, Salescast
anticipates inventory problems; however it assumes that inventory levels are correct. For retailers, Salescast and Shelfcheck are complementary. However,
Shelfcheck only applies to stores, while Salescast applies to both stores and warehouses.Big Picture
Studies (
*) show that on average, 5 to 10% of the products offered in grocery stores are unavailable at any given time, a situation which has not improved for 2 decades. Products that are not available to the consumer on the dedicated shelf space are referred to as being
out-of-shelf (OOS). Monitoring OOS is the cornerstone of
on-shelf availability optimization. Shelfcheck, a webapp by Lokad, delivers the OOS metrics you need along with alerts to optimize your on-shelf availability.
Through
full historical analysis of past OOS issues, your company gains immediate visibility on the OOS problem. Let store managers benchmark product availability over time, or benchmark products against each other; not mentioning benchmarking stores directly.
Consume
OOS alerts anywhere, even on your smartphone. Shelfcheck is optimized for low-bandwidth and low-resolution device.
Get the
most accurate OOS analysis on the market, as Lokad manages all demand patterns on your behalf: seasonality, trend, promotion, product life cycle... No statistical skills are required.
Pricing is by monthly subscription and
depends on the sales volumes of your store. Limited setup cost. No termination cost. When you stop using Shelfcheck, we stop charging. No questions asked.
Achieve
complete process automation with Shelfcheck. Pull data from your current IT systems, and get results through web reports, or directly injected within you existing business apps. Ongoing maintenance of Shelfcheck requires near-zero manpower.
The detection of out-of-shelf products
Many root causes exist for out-of-shelf (OOS) issues:
- Product has not been ordered in proper quantities.
- Product is not at the right place and customers can't find it.
- Product has been stolen and incorrect electronic inventor level prevents reordering.
- Packaging is damaged and customers put back product after having a close look.
- ...
Whenever a product is not available on the shelf,
sales are lost. Studies (
*) show that clients will substitute with available products only 1/3 of the time. OOS situations causes a loss of sales and generates
client frustration which turns into loss of loyalty. OOS drive your clients to your competition.
How does Shelfcheck work?
Shelfcheck leverages an
indirect measurement method. Instead of trying to assess the physical state of the shelf itself, we identify an OOS by analyzing and identifying the impact on store sales. For example, if sales of a given product drop to a level that is significantly lower than expected (or 'normal'), then the product is very likely to suffer an OOS issue. A multitude of such patterns and signals is monitored and identified by Shelfcheck, which in combination give a clear picture that is reported by Shelfcheck.
The key ingredient for such an OOS analytics technology to work is a state-of-art
forecasting technology. Indeed, OOS are identified as a divergence between
observed sales and
expected sales (aka forecasts). Our technology combines a
vast library of statistical models, an extensive model selection mechanism, a multiple time-series modeling approach and a
native cloud computing infrastructure to power such a processing-intensive solution. As a result, all
common patterns are taken into account: seasonality, trend, day of the week, promotions, product life-cycle and much more.
Does Shelfcheck apply to your company?
If you have shelves holding fast rotating inventories, yes, it does. Food retailers are the
archetype of businesses who benefit the most from Shelfcheck; other segments such as do-it-yourself (DIY) stores, for example, benefit also from this technology greatly as well. As a rule of thumb, our technology creates the most value-add when at least a
small percentage of products get sold once a day on average.
Then, part of the problem is that without an OOS metering system such as Shelfcheck, retailers are
blind when it comes to OOS issues. Since Shelfcheck comes without any upfront investment, it offers you the opportunity to
benchmark your out-of-shelf rates for a very modest cost that requires neither management bandwidth nor intensive manpower from store employees. Doing a quick benchmark with Shelfcheck is the best way to assess where your company stands.
Why choose Shelfcheck over alternatives or competitors?
Improving shelf availability is a century old fight. Yet, traditional solutions suffer from significant drawbacks:
- Direct shelf control is extremely labor intensive.
- RFID remains too expensive for most retail segments.
- Rule-based alters are too inaccurate to be of practical use, except for super top sellers.
The technological challenge behind Shelfcheck is the
quality of OOS metrics. Indeed, it's easy to design a software that churns out a truckload of OOS alerts, but the time of the store staff is precious - and expensive - and employees can't waste their energy chasing
false-positive alerts, or
phantom problems advertized by the software, but having no real counterparts within the store.
By leveraging a superior
forecasting technology, Shelfcheck delivers OOS metrics with both unprecedented sensibility and precision, the two metrics that matter most when it comes to OOS analytics.
Sensibility vs Precision
An indirect OOS detector such as Shelfcheck (or any of our competitors) relies on the analysis of the
divergence between observed sales and expected sales. Yet, no matter the quality of the statistical analysis, the possibility of very random (unpredictable) market fluctuations is unavoidable. Hence, OOS monitoring cannot be 100% accurate. Yet, not being perfect does not imply being worthless.
There are two key concepts when dealing with OOS analytics:
- The sensibility that represents the percentage of true OOS captured by the software.
- The precision that represents the ratio of true OOS within all OOS reported by the software.
Neither 100% sensibility nor 100% precision is possible; or rather 100% sensibility implies 0% precision (resp. 100% precision implies 0% sensibility).
Through a superior
forecasting technology, it is possible to improve both sensibility and precision. Then, on top of the forecasting layer, Shelfcheck offers a
fine-tuned trade-off between sensibility and precision that allows the adjustment of systems to the specific application and customer.
Features: OOS analytics and OOS alerts
Shelfcheck delivers to two distinct sets of features:
- For managers: an analyzer engine providing transparency and benchmarks a crossed the network by reporting all OOS issues detected down to a daily aggregation per product.
- For staff: an alert engine that delivers accurately prioritized alerts concerning OOS issues.
The data, both OOS analysis and OOS alerts can be
consumed through the web from Shelfcheck itself, or
injected into your existing IT systems.
Analyzer data
The analyzer engine produces data that can be leveraged by store managers to
gain insights on OOS issues, and improve overall practices. The graphic below illustrates the data collected by Shelfcheck: missed sales quantities are reported per store, per product and per day. When no OOS are detected, cells are left blank.
Then, the very same data is also reported in the local currency.
The data included in those two graphics is
fictitious, but you might notice that the OOS quantities reported for the first day of the OOS issue is typically lower than the one of following days. We are observing here a very common pattern: OOS issues start with a
partial OOS with some inventory left that gets depleted during the course of the day. The partial OOS is then followed by a
full OOS with no inventory left right from the beginning of the day. The OOS lasts until a replenishment is made.
Naturally, this
data can be aggregated at higher levels, such as per product category or family. The intent of Shelfcheck is to give you access to the raw OOS data at the finest level possible; yet Shelfcheck is not a replacement of a Business Intelligence portal.
Many operations can be accomplished with this data:
- Quantify total revenue and profit loss caused by OOS.
- Identify OOS outliers, such as poorly available products.
- Benchmark one store vs another store.
- Benchmark one product vs another product.
- Refine in-store safety stocks.
- Identify poor performing suppliers.
- ...
Complementary apps
Shelfcheck is not intended to be a
one size fits all system. First, Shelfcheck relies on the existing IT system to import store sales data. Second, the raw data produced through OOS analytics can be made available to other 3rd party apps.
In particular, Shelfcheck data can be injected into an existing
Business Intelligence portal in order to reduce the end-user learning efforts to a negligible level.
Pricing
We charge on a
monthly basis based on the number of stores, with a store price that increases along with the size of the store. Please see for details our
pricing page .
Shelfcheck is delivered as a software-as-a-service, there is no further cost in addition to the Lokad fee (i.e. no need for hardware, software, hiring, training, maintenance).
Getting started
Shelfcheck is still in limited beta stage. In order to get started with the product you will need to
get in contact with us. Just drop an email to
contact@lokad.com.
We have adopted a flexible architecture that let us
plug an arbitrary data stream into Shelfcheck. In practice, data is transferred either through flat files or through SQL databases.
Data required for Shelfcheck:
- (required) Daily quantities sold per store per product, with daily rolling updates. The more depth in the history, the better the analytics. Those quantities are fundamental for the Shelfcheck to perform the OOS analysis.
- (optional) Ticket details with prices. The fine-grained listing including basket data (instead of daily aggregates) can be leveraged by Shelfcheck to position more accurately the starting time within the day of an OOS for top sellers. Then, the prices let us refine the prioritization of the alerts.
- (required) Product catalog listing, with families, sub-families. Monthly or quarterly updates are recommended. Product descriptions are essential to produce intelligible reports.
- (optional) The list of must-have products, or some equivalent short-list of highly desirable products. Those listing can be used by Shelfcheck to produce better prioritization of the alerts.
- (optional) Daily replenishments per store per products, with daily rolling updates. Replenishment quantities can used by Shelfcheck to increase both sensibility and precision of OOS alerts issued by Shelfcheck.
- (optional) Promotion flags per store per products, with daily rolling updates. Promotions typically cannibalizes non-promoted products, which can lead to a decreased precision of the OOS alerts (low sales being explained by the promotion, not by an OOS). Shelfcheck can leverage promotion data to refine demand forecasts.
The first phase typically consists by establishing a
proof-of-concept (POC) for a single store. Depending on your resources to extract more data, we are also happy to do a POC for the entire retail network if the data extraction is within your operational possibilities. From this POC, you will gain access to an extensive
historical OOS analysis per store, per product and per day.
Security notes
From a legal viewpoint,
we treat every single piece of data exposed to Lokad by prospects and clients as if the data were under a strict NDA (Non Disclosure Agreement). The data are and will remain the property of your company. Security is the second most important concern for Lokad (forecasting accuracy being the first). If needed, we will gladly sign an NDA with your company on simple request. Just
email us your template.
From a technical viewpoint, Lokad is hosted on Windows Azure, the cloud computing infrastructure from
Microsoft, a computing environment that we believe to be about one of the most (if not the most) secure on the market. By relying on Windows Azure, we benefit from all the ongoing security efforts and monitoring that Microsoft is applying to its own computing infrastructure.
References
- Improving On-Shelf Availability for Retail Supply Chains Requires the Balance of Process and Technology, Gartner, 26 May 2011
- Optimal Shelf Availability, Increasing shopper satisfaction at the moment of truth, Roland Berger Consultants, 2003 download report
- Retail Out of Stocks - A Worldwide Examination of Extent, Causes, and Consumer Responses, Thomas W. Gruen (University of Colorado), Daniel S. Corsten (University of St. Gallen), Sundar Bahradwaj (Emory University), 19 May 2002 download report