Q: After Signing up on IRIS, is it safe to start uploading data as inputs right away? Is there any NDA (Non-disclosure agreement) or EULA or any other agreement which ensures that my data is secure and will not be accessible to any third party?
A: During the sign-up process, before you submit your details, you get to see terms and conditions on the page which includes a EULA (end-user license agreement). This agreement has all clauses necessary to ensure that your data is secure and you can start your uploads on IRIS without any hesitation. You can refer to the agreement on the following link – EULA.
Q: How will projects be created, at the company level or for different sales channels?
A: The concept of projects has been created to provide flexibility to the organization using IRIS. Projects can be created for different brands, different channels, regions, a combination of these or any other way based on how the roles and responsibilities of the users are in the organization. The only caveat here is that by creating a lot of projects, an organization may lose the advantage of analyzing data at a central/macro level.
Q: Are primary sales good enough for the tool or is it important to upload the secondary sales?
What if we do not have secondary sales for some of our channels?
A: It is very important to have secondary sales as an input in IRIS as that gives the actual performance and is the correct input to use for predicting demand mix. Primary sales can be misleading and do not take into consideration the secondary sales that could not happen (dead stock, returns, etc.) which directly impacts the accuracy of IRIS outputs.
Q: Do you incorporate customer returns? Is there a size that is being returned a lot? We might
reduce producing it
A: In offline, the customer physically tried, tested a product before buying. The no. of returns are relatively low. Hence the sales data is far more meaningful for analysis as opposed to returns data. In online, the customer can return the product or multiple reasons –
- The physical product was not the same as shown in the catalog.
- It was not of the right size
- The product got damaged during transit
- Payment problems
- The customer was not available during delivery
All these reasons do not give any clue as to whether the product was desirable or not, from a consumer PoV. Hence the system algorithms focus on what consumers opted to buy as opposed to what they returned. If still needed, it is recommended to subtract returns from sales offline before uploading the transactions.
Q: How much transaction data is required for analysis?
A: It is good to have at least a year’s data as it covers the sale trends across all months (covering seasonality) and is helpful to choose different analysis periods from the same easily basis requirements.
Q: We do not have uniform SKU codes, there are multiple SKUs of the same product. Can IRIS handle this? Is there an SKU mapping feature?
A: There is no SKU mapping feature in IRIS (It’s there in IRIS enterprise), so it has to be made sure that each product’s SKU code is unique. Else the user needs to manually map the SKUs to a single SKU offline before upload the inputs.
Q: How is IRIS better than Microsoft Excel?
A: IRIS works on a set of patent-pending algorithms in the backend which have been made after years of research on the planning and distribution done by merchandisers. This was achieved by working directly with multiple brands’ merchandising team on IRIS enterprise and identifying key areas of improving the accuracy of outputs and automation. IRIS incorporates all those learning and has been made with the vision of making life as simple for merchandisers as possible with high accuracy and speed. Not to mention, all calculations done in IRIS are at a very granular level – store + product attribute group, which becomes a challenge when done manually in Excel or a similar tool, esp. when the number of stores or product attribute groups are high.
Q: Can you integrate to the partners via API or FTP for data to automatically flow in without the need to upload?
A: No, as of now IRIS does not support system integrations and inputs must be uploaded directly on the tool before usage.
Q: Why are TSV files used instead of CSV?
A: CSV files run into problems when there are special characters like comma (,) or quotes (“) in the data. TSV files can handle special characters very gracefully.
Q: In which format should the dates be uploaded in sales data?
A: IRIS accepts dates uploaded in “yyyy-mm-dd”. In most cases, the format is set to “dd-mm-yyyy” and might come the same way on downloading the templates/demo data before upload. So it is recommended to change these dates to the correct upload format of “yyyy-mm-dd”. This can be done easily by using TEXT(“date”,”yyyy-mm-dd”) Excel formula.
Q: Do you consider inventory? How do you correct for availability/exposure?
A: No, IRIS does not take daily historical inventory as an input. Availability/exposure is corrected in the algorithm via a statistical technique. However, store and warehouse inventory snapshots are taken as input before running distribution.
Q: Are algorithms the same for all kinds of industries?
A: Algorithms have been made keeping the apparel industry in focus but have worked smoothly for other sectors like cosmetics, accessories, footwear, etc.
Q: How does IRIS process (clean) the raw data to predict true (correct) demand from raw sales? How are liquidations (discounted sales) and unavailability (of inventory) accounted for while predicting demand?
A: The first step before consuming any data for analysis in IRIS is that of data processing called the “Sales data cleanup”. In this step basis the limit defined by the user as an input, IRIS cleans up the most liquidated (discounted) sales of each attribute group in respective stores. While doing so, IRIS also modifies the sales data with smart algorithms to normalize sales drops due to unavailability during the sales period selected.
Q: I see lesser “Percentage of Revenue” cleaned. What can be the reason for this?
A: Say you gave 10% value for the Max Liquidation Cleanup Cutoff %, but now you see only 8% as Percentage of Revenue Cleaned. The reason for this is that the system intelligently buckets similarly discounted data, and either includes all or none (as they are similar). This is illustrated below
So, in the above example, there is not much difference really in giving a 35% discount or a 40% discount (they are similarly discounted). So, either all such sales should be either cleaned up together or not cleaned up. In order not to breach the 10% criteria, the system chooses not to clean up such data. Hence you would see only 8% of revenue cleaned up.
Q: How do you plan to adopt this forecasting tool for categories other than Fashion – FMCG, White Goods, Accessories & Cosmetics?
- IRIS will work perfectly well for categories other than Fashion.
- We have tested IRIS forecasting for FMCG, White goods, Accessories & Cosmetics categories along with Fashion and the results have been satisfactory.
Q: Why in Iris line of product we use FTP to transfer data rather than have API integration?
A: Increff Merchandising solution is an analytical platform that takes large volumes of data to run its algorithms and deliver intelligent output. It is built for small and large clients whose daily data input can run into millions of rows (1000 of stores).
The best practice for transferring large volumes of data is FTP or file in a bucket. Normally, API has the best use case with transactional systems i.e. when you have to give a response based on the data provided. For example, an order is placed on the website and if there is a service provider that can give tax computed receipts, then it has to be generated in real-time (response).
Increff Merchandising solution is a predictive system where one day’s data may not be sufficient to give insights. For example, if you are using it for OTBs there can be other constraints like MOQ, etc. which need to be considered when making a decision. The solution consumes daily data to analyze the best use case for running your ARS (Automatic Replenishment System), and for any other analysis, the solution consumes large volumes of data, of weeks to months to years, to accurately predict the optimum demand.
Q: How is IRIS going to work for online channels? Is it even useful on that front?
A: IRIS can be used for online channels as well. Each channel can be created as a store and sales transactions can be uploaded against the same. The tool does its job of identifying top sellers, recommending size rations, projecting sales mix, splitting revenue targets, and providing sales projections. If required, it can also do replenishment for each channel.
Q: How does IRIS help in selection for scenarios where orders are placed by external buyers? For example, Shoppers Stop stores where buyers over there decide what a brand gets to keep.
A: IRIS identifies top sellers for each partner and recommends the correct mix of attribute groups in which the merchandise should be bought for each store. Since this is a data-driven insight, it is important to pass on the same to the buyers when they place orders, in order to influence their decision making to improve sales which is beneficial for both the parties.
Q: Can more attributes be added to get the idea of performance and direction on what to produce
A: IRIS currently supports brand, sub-brand, gender and price bucket as the only levels of attributes, but these can be used smartly to incorporate more attributes, for example in Denims, Fit can be incorporated by creating Slim-fit Jeans as a subcategory within Denims. If there is a requirement for more detailed accuracy – including more attributes, it is recommended to move to IRIS enterprise which is capable of handling up to 17 attributes. Below is a video on how to do it –
Q: Business does not have a concept of NOOS, is it mandatory to have NOOS or can we get predictions across fashion clusters only?
A: The tool can identify top sellers and it is recommended to reorder the same, but it is completely a user’s call to take advantage of this insight.
Q: Can IRIS handle separate forecasting/planning for core and fashion styles? Here the core products are the ones defined by brands.
A: IRIS can handle this problem in two ways. Firstly, if the requirement is the get the overall demand for core clusters and fashion clusters separately, the user can concatenate the Gender attribute with the “Core/Fashion” term. For e.g. Men’s Core and Men’s fashion can be two separate genders. This way, their clustering will be different. On the second hand, if the requirement is to predict the demand of individual core styles, then it is recommended to put the core style code name as one of the attributes so that single style clusters are created for each core style and all analysis is done – outputs given at a style level.
Q: What if we changed the size measurements in between? For example, the size measurement is now an S size measurement?
A: In cases like these if the changes in size measurements are linkable from size to size, it is recommended to update the historical sizes as new sizes. If that is not possible, then there is a flexibility of selecting a specific analysis period where a user can choose a period where the new sizes are selling.
Q: How are exit sized determined by IRIS?
A: During the data processing part, IRIS asks for user’s input as a benchmark to exit certain sizes in respective stores where their sales contribution has been below the benchmark. For e.g, if the benchmark is 2%, IRIS will recommend most sizes contributing below this 2% benchmark of sales at the product group level (Brand-category-gender) for that store.
Q: Different brands and different categories have different pricing strategies (across EOSS/full price/online discounting). Then how does the liquidation strategy work?
A: IRIS works at an attribute group + store level which is more levels below the brand – category level. So, the liquidation cleanup of historical data analyzed also happens at those levels and the price bucket mix within a brand – category – sub-category is also given as an output that can be used to correct pricing strategies directionally. As far as different periods are concerned, IRIS gives the flexibility to the user to choose the analysis period for every output, so a person can choose a historical EOSS period for
future EOSS prediction and likewise for full price.
Q: Does this tool forecast demand for a new category to be launched?
A: No, the tool can only forecast for categories present historically.
Q: For how many months in the future should the AOP be uploaded?
A: For as many months the sales mix projection (revenue split) is required.
Q: How to change the AOP input on the basis of a Brand being seasonal (Like to Like historical period for prediction) or id the brand is non-seasonal (Recent sales to be considered for prediction)?
A: Below is a video to explain the same in detail –
Q: Tool is quite relevant for brands that have their own EBOs as they can use it for distribution, as they will have complete control on what stocks to send to stores, and if they are doing it manually.
A: Yes, for one’s own store control on execution is 100% but even for external partners, IRIS is useful to keep a check and validate the allocations and replenishment. It has been helpful in the past to highlight gaps and course correct the stock situation even at partner stores of brands on IRIS.
Q: How does IRIS maintain correct merchandise assortment at stores while allocating/replenishing styles based on individual style performance? Is it possible that the last style allocated for a store-category due to planogram gap constraint is a lower-ranked style instead of another available higher-ranked style?
A: This is one of the USPs of IRIS where IRIS takes the planogram input from the user at a store-category level and intelligently breaks it down in the backend in the right assortment of product attribute groups. Due to this, the planogram constraints work at a level below category. This ensures that if there are two styles of different sales performance, but one of them is more important to complete the recommended assortment, then that style is given priority for allocation. This check is dropped if the overall planogram input is still not met and the tool ensures that the stocks are ultimately fulfilled at a store-category level planogram.
Q: How to avoid the allocation of old season styles?
A: Simply remove them from whstock (Warehouse Stock) file, before uploading it
Q: How to prevent broken styles during allocation?
A: The system tries its best to do allocation in a manner where only non-broken styles go first to the stores. If the stores remain empty, then it tries to allocate styles which have good quantity left in the warehouse but have sizes missing. If you do not want your first allocation to contain any broken styles, you can simply remove these from the output.
Q: How can I create a planogram for online channels?
A: As online channels do not have any display constraints, creating a planogram is not based on physical constraint but based on the forward cover that the business wants to carry. Since IRIS gives you projected sales quantity for upcoming months after breaking down your revenue targets, you can take that as a direct output from the tool and sum up x no.of future months’ projected sales quantity you need to take as a cover and use it as the planogram quantity. In case you want to work on the backward cover, i.e., basis recent historical sales, you can simply take the historical x duration sales as your planogram quantity constraint. So, for example, a subcategory of polo T-shirt, if you project to sell 300 pieces per month on the channel and you want to keep a forward cover of 3 months inventory with the channel. Then your planogram will be 300*3 = 900 pieces.
As a brand, if you plan to keep 50 pieces in each style, this will become the size set for a polo T-shirt.
Q: Does the tool support multi-warehouse set-up? For example, the demand of Haryana/Karnataka warehouses in terms of taste of customers/size ratios might be different. Can this tool help understand the demand across different warehouses?
A: IRIS does not support multi-warehouse distribution and warehouse to warehouse transfers. IRIS enterprise can handle a multi-warehouse to store distribution but that too has limitations on the warehouse to warehouse transfers. As far as understanding demand patterns (attribute group mix and size) is concerned, that can be easily provided by IRIS, given that the inputs are also provided at the same level.
Q: Does the tool ensure that size continuity is maintained while allocation and broken styles are not allocated to the stores?
A: Yes, the tool automatically identifies through set logic if a style is healthy enough to be allocated and a minimum x number of key sizes are present and continuous.
Q: While distributing is there a feature for coordinates (sending a top/a skirt together)?
A: No, such a feature is not there in the tool as it does all the distribution based purely on the merit of individual styles and recommended assortment for each store.
Q: Does IRIS ever breach planograms in distribution?
A: No, IRIS will never exceed a store-category level planogram constraint.
Q: If we want to do a new allocation basis historical period and replenishment basis in-season, can we have two different analysis periods?
A: Yes, you have to select the analysis period in the distribution for new allocation and replenishment period
for in-season replenishment before running distribution.
Q: Do you consider the expiry date of the products? Does it handle the FIFO distribution?
A: No, IRIS does not take expiry dates of products as an input
Q: Will this work for primary replenishment (distributor to channel partners) and secondary replenishment (partners to end consumers)?
A: IRIS will work best where secondary sales data and store inventory is available. In most cases of distribution, this information is difficult to get – not to mention the stock visibility at the distributor’s warehouse is another complication which as of now IRIS does not handle.
Q: instead of the number of pieces, can we input DOH cover in a planogram?
A: no, currently the tool takes only quantity in pieces as the capacity constraint in the planogram. A user can convert the DOH cover in pieces and upload them before running distribution. In case the user needs the tool to help create planogram basis DOH input, then that feature is available in the Enterprise version of IRIS.
Q: Apart from product outputs, are there reports where we can see the brand’s own data?
A: yes, there is a separate BI module in the product where a set of reports are there for analyzing the brands’ actual sales data. These reports have been made keeping in mind the key insights a user would like to draw from the data uploaded on IRIS and cover most possible views of importance. In case a user feels there is a need for an additional view, he/she can reach out to IRIS customer support and if the team finds it as an important and doable view, it can be added.
Q: Is IRIS/IRIS Enterprise suited for SMB (Small Medium Businesses)? Do you have pricing for SMB customers?
- IRIS is suited for SMB customers and IRIS-Enterprise is suited for Large scale businesses.
- IRIS Pricing – Depends upon business size and is negotiable. Connect with us for details
Q: What about the features of IRIS and IRIS Enterprise license – a comparison chart will help
A: IRIS v/s IRIS Enterprise Comparison – Click Here
Q: When we buy a license say for eg. Fashion business, then can we use it for all businesses – Fashion all brands, FMCG, others or for every business we need to buy a separate license?
- IRIS offers a unique concept of Project. You have the flexibility to create a different project within the same license for every business – Fashion or non-Fashion, thereby handling all brands under the same license.
Q: Do you have a performance-based licensing model wherein we pay you basis key B&M KPI’s like inventory turns/aging/stock availability at the right store at the right time?
A: No, IRIS is a pure-play self usable SaaS model with standard pricing, connect with us for more details. You can use it for topseller management, planning, distribution, regional utilization, BI & Return Management. At this cost, you can work on unlimited data, unlimited sales channels, brands and styles. There are no other hidden costs like onboarding or support costs. On top of this we keep on upgrading our product and release new features almost every 2 weeks, this comes as an additional benefit with no extra costs involved.