Home »
Salescast » Here
Does Salescast apply to your company?
If you have inventory, yes it does. If your company has inventory, then someone, somewhere, is already producing forecasts, because inventory is nothing but anticipated demand. And more often than not, the person doing it doesn't have the word
forecast in the job title. Nevertheless, the
level of inventory optimization that your company is able to achieve is directly related to the accuracy of those forecasts.
Big vs Small
Salescast has been built to produce sales forecasts for any company that needs to anticipate demand in order to optimize inventory and other resources. Salescast is well suited for
retail, wholesale, eCommerce and manufacturing. Our largest customers are large retailers with thousands of stores, and our smallest customer is an eCommerce with 2 employees.
If you're a huge company, Salescast will scale on demand to hundreds of points of sales or inventory locations. This scalability is only possible because Salescast has been
natively built for cloud computing.
If you're a tiny company, it means that Salescast will save you a lot of time and effort, producing forecasts that would otherwise require statistical skills way beyond your resources. Our forecasts are delivered as a service, we do not sell bare forecasting toolkits, and you don't have to put any extra effort to get accurate forecasts.
Yet, unlike nearly all other forecasting tools offered by competitors, Salescast requires
zero statistical skills and next to zero manpower. The product has been built from scratch will
full automation and
unlimited scalability in mind. We aim at delivering most accurate forecasts, which will enable your company to
reduce overstock and reduce stock-outs at the same time.
Fresh food vs Car parts
As our technology is completely statistical,
Lokad does not rely on domain expertise to produce its forecasts. As a result, Salescast applies indifferently to fresh foods, fashion consumer goods, automotive parts, or any other markets. We are pretty much market agnostic, as long some historical data exists, and that overall sales numbers are compatible with a statistical approach - i.e. if your company is signing about 2 large contracts per year, then, indeed, statistics do not apply.
Nevertheless, we are very
capable of dealing with the specifics of your business. For example, bulk orders, long lead time, very spiky demand, stock-outs, ... as a few of the many factors natively accounted for by our forecasting technology. Through Lokad, you're getting access to a forecasting technology that has been polished on many similar businesses already. Indeed, we treat each new client dataset as an opportunity to learn more about demand modeling and forecasting, and consequently improve our technology.
Zoom: Large retail networks, what it takes to be scalable
The scalability challenge: A particular forecasting challenge for large retail networks is the sheer amount of products in their portfolio, multiplied by thousands of points of sale (PoS), which results in millions of SKUs to be forecasted. Nearly all forecasting solutions on the market claim to be
scalable, but in our experience very few live up to this claim when dealing with large retail networks.
Situations that reveal
non-scalable forecasting solutions include:
- You can't get a full-scale benchmark (or Proof of Concept) within a few weeks at a minimal cost. If the solution does not offer the possibility to make a simple quantitative assessment over an arbitrarily large amount of data with a minimum of efforts, then, transitioning the solution toward production is going to be a titanic undertaking.
- The manpower needed to support the forecasting solution grows steadily with an increase in the number of sites, products or sales patterns. This is not scalable, because following this pattern, advance forecasting at the point of sale level is going to require nothing short of a small army of people to work out.
- The system needs a lot of simplifications and high aggregation level to work. Systems that can only forecast on a warehouse or product group level (which then need to be broken down) have clearly surpassed their scalability. These simplifications come at the cost of significantly reduced accuracy.
- In-house hardware is required to setup the solution. This is not scalable either, because forecasting is an intermittent need: forecasts are typically produced once a day, you certainly don't want a computation running for 24h to complete. If you rely on in-house hardware, the hardware will be sitting around unused more than 90% of the time.
- The accuracy does not naturally improve as the amount of data grows. If feeding the forecasting system with data of hundreds of POS does not improve the overall accuracy compared to warehouse-level forecasting, then the solution is not scalable. Indeed, with statistics, the more data, the better the analysis. The scalability of the forecasting solution should reflect this statistical fact.
Lokad is compliant with this scalability real life test: We believe we offer one of the few (if not the only) scalable forecasting solutions on the market that really embrace the scale of large retail networks.
Lokad delivers robotized POS-level forecasts as well as warehouse-level forecasts. Doubling the number of POS or switching from weekly to daily forecasts is done at the touch of a button and the required manpower is close to zero.
The promotions and product launch challenge: Retailers do not only have to deal with a huge amount of SKUs, but also with a dynamically changing product portfolio and repeating steep rises in demand due to promotions. We do not dismiss these tough forecasting cases and provide completely automated forecasts of promotions and product launches.
Note that automated forecasts
do not prevent the retailer from performing
manual overrides where it makes sense.
Zoom: eCommerce, a need for speed
Few markets are as fast paced as eCommerce. Technology evolves rapidly, competition increases as the market matures, and strong growth often needs to be managed. Yesterday it was social networks, today it's mobile commerce, and who knows what will be the next big thing to shuffle all cards again. We believe that Salescast has a number of characteristics that make it particularly
suitable for eCommerce:
The shortest delivery time of the market: Lokad is not only about delivering the most accurate forecasts of the market, it's also about rapid implementation. Because your sales data is already centralized in an SQL database, Salescast can deliver the first batch of forecasts literally within hours. All it takes is to expose your sales data following our guidelines (see
Intermediate SQL schema) which typically requires about 2 or 3h of work from a database administrator familiar with your company IT system. Classical demand forecasting tools with their typical 12-months PoC-to-production timeframe are simply
unfit for eCommerce. By the time the project completes, everything has changed once more, and the hard-earned forecasting setup needs further extensive revision.
Strictly on-demand: No upfront fee and a pay-as-you-go pricing. If your company is undergoing massive growth, Lokad scales up to match your needs. Yet, the other way around, if your company takes a hit or simply decides to refocus on fewer products, the Lokad cost scales down as well. If the market proves tough, unlike many competitors, Lokad will not become an oversized investment, dragging your company down even further.
Built for large product catalogues: Our technology has been designed with many characteristics of eCommerce in mind. For example, we typically observe that the number of items sold by an eCommerce within one month is roughly equal to the number of items in its catalog. This means, that the typical eCommerce sells on average each product only once per month, and many products follow an
intermittent sales pattern. Most sales forecasting tools struggle with such patterns. Of course, if a product is sold once a year, then there is little to be achieved through sales forecasting. Nevertheless Lokad is able to handle sales volumes substantially lower than what is typically required by competitive tools.
As a final note, Salescast has been designed
with full automation in mind. eCommerce is often about scalability, process optimization and cost. Automating the otherwise tedious job of forecasting sales is a very good way to increase efficiency and decrease costs. With Salescast you will not only benefit from full automation and Microsoft Excel forecast reports, but also from direct SQL integration with Salescast that can inject its forecasts directly into an SQL database of your choice. This lets you further automate your business feeding the workflows and business reports that best fit your specific needs.
Zoom: Manufacturing, moving beyond Excel
Most probably, in your business, as far anyone can remember, demand planning has always been a challenge. And over the last two decades Microsoft Excel has been the King to produce those numbers. We believe Salescast represents the next-generation approach that will bring your business to a whole new level as far forecasting is concerned.
We are not dismissing Excel though, it remains excellent to visualize and share data. Actually, Salescast delivers Excel reports. Nevertheless, we also believe that Excel, by itself, offers very little to support the recurrent delivery of accurate demand forecasts.
Here are a few problems that we frequently observe with Excel-based forecasts:
- Forecasts are fragmented, typically over many spreadsheets. Furthermore, those spreadsheets are loosely updated by many people, possibly in over multiple production sites.
- Forecasts involve a lot of untold know-how. Some people are producing much more accurate forecasts than others. Yet, the training process to get better at forecasting is fuzzy at best.
- Teams have hard time keeping up with the weekly or monthly updates. As a result, old forecasts are often copied and pasted as new forecasts whenever there is too much immediate pressure.
- Teams also lack confidence in their own forecasts. As a result, people are frequently unwilling to share their forecasts with anyone but their immediate management, which requires these forecasts in the first place. This creates friction between sales, production, supply chain, management and suppliers.
However, salescast is no silver bullet that will make all your supply chain problems go away. As a matter of fact, Salescast only focuses on providing a
much better alternative to Excel for demand planning.
As we always suggest benchmarking Lokad, we observe that our forecasts are typically much more accurate
on average compared to the forecasts produced through Excel. Here it's really the
average situation that matters. Indeed, we don't question the fact that when a business expert spends 2 hours on making a forecast, this number happens to be more accurate than the one offered by Lokad. Yet, as far as we can tell, typical Excel forecasting processes involve no more than a few seconds of attention for each number produced.
Besides, through complete automation, your teams benefits from forecasts that are delivered right on time,
any time you need them. Since forecasting process is more reliable, there is much less resistance in letting more people
in your company
access those forecasts. And consequently a
lot more people can take advantage of those numbers to
optimize their processes.