Omnichannel starts with one stock number

You check a dress on your phone, see "in stock", walk to the shop and are told the size has gone. To the customer, I would say, that is not the website's fault or the shop's fault; it is the brand's. This post asks one question: for a retailer to keep the promise "what you see is true wherever you look", what must it know first?

How does omnichannel differ from multichannel?

Verhoef, Kannan and Inman (2015) open their special issue of the Journal of Retailing by noting that the move to omnichannel retailing had not yet been properly conceptualised, despite its growing importance in practice (Verhoef et al., 2015). They define omnichannel management as the synergetic management of the numerous available channels and customer touchpoints, so that the customer experience across channels and the performance over channels are optimised. They add that channels interact and are used simultaneously (Verhoef et al., 2015).

Table 1 in the paper sets multichannel beside omnichannel. Multichannel covers interactive channels only: store, website and direct marketing by catalogue. Each channel is separate with no overlap, and is managed against its own objectives, such as sales or experience per channel. Omnichannel adds mobile channels, apps and social media, plus touchpoints in mass media such as TV, radio and print and contact between customers, and adds the brand to the focus alongside the retail channel. The channels are integrated so that the retail experience runs on without breaks, and the objectives are the overall customer experience and total sales across channels (Verhoef et al., 2015).

The figure shows what a customer does: researches on mobile, chooses online, adds to the basket in store, collects through a physical channel, returns in a shop. Verhoef and colleagues say shoppers move between desktop, laptop and mobile devices, and may inspect goods in a shop while hunting for a price on their phone (showrooming), or search online and buy in a shop (webrooming) (Verhoef et al., 2015).

Why do I think stock is the bottleneck?

This is my reasoning. I picked four promises that turn up often in omnichannel retail: "see what is in the shop", "order online, collect in store" (Click&Collect), "order from the shop and have it delivered" and "buy online, return in a shop". All four require the firm to know exactly what is where at the moment the customer asks. A wrong stock figure does not just spoil one order; it breaks the shared promise.

Inditex's report links RFID to stock in this way. It says RFID "allows us to track every one of our garments at all times, from when they arrive at our central platforms until they are sold", and that RFID has enabled the online-offline stock management system, so customers have better access to the products they want wherever they are. The report also lists accompanying services: Click&Collect, ordering online from a shop for delivery to a pick-up point or the shop of the customer's choice, and returning or exchanging online purchases in the group's shops. It does not say RFID is a condition of each service (Inditex, 2019). By year-end the integrated system was operating in 49 markets where Zara has an online and physical presence, and the goal was to have it in all brands by 2020 (Inditex, 2019). See also the Inditex case.

Which figures support it, and what can they tell us?

The report gives several figures around the system. RFID is fully deployed at Zara and, according to the report's summary, at Massimo Dutti, Pull&Bear and Uterqüe; Inditex has invested over EUR 2 billion in technology over six years; same-day delivery covers 12 metropolitan areas and next-day delivery eight markets (Inditex, 2019). Online sales grew 27 per cent to EUR 3.2 billion, 12 per cent of net sales (14 per cent in markets with online sales) (Inditex, 2019). Shops receive new merchandise twice a week; the average time from receipt at the distribution centres to delivery in shops is 36 hours in Europe and at most 48 hours in the Americas and Asia (Inditex, 2019).

None of this answers the question "does unified stock pay for itself?". There are three gaps. First, the report describes the inventory efficiency as without precedent and offers no measure (Inditex, 2019). Second, there is no figure for orders fulfilled from shop stock, for Click&Collect orders or for cancellations caused by stock-outs. Third, net sales in 2018 were EUR 26.15 billion, so online was about one-eighth of the total, and the report does not isolate which sales the stock system produced (Inditex, 2019).

What does academic research say about payback?

Less than people assume. Verhoef and colleagues summarise three studies in the special issue. Cao and Li analyse secondary data on US retailers and find that integrating multiple channels is positively related to sales performance, with some contingencies moderating the effect. Herhausen and colleagues run three experiments and find that online-offline integration produces channel synergies, not cannibalisation (Verhoef et al., 2015). Pauwels and Neslin study a retailer that already had online and catalogue channels and then opened a shop: the shop cannibalised catalogue sales, had much less effect on internet sales and raised returns and exchanges, but purchase frequency across channels rose, so total revenue rose; data limits meant profit could not be studied (Verhoef et al., 2015). That is a study of adding a channel, not of integrating channels.

At the same time, the editors write that integration and harmonisation across channels are "generally considered as good, but is this always good?". Among the research questions they propose are two: to what extent retailers should really strive for integration, and whether a smooth experience across touchpoints actually produces stronger performance (Verhoef et al., 2015). On showrooming, a survey of salespeople finds that perceiving it is negatively related to their performance, and that cross-selling and coping skills may weaken that relationship (Verhoef et al., 2015).

Where should a small shop start?

The example below is hypothetical. The shop's name and every detail are my invention.

A hypothetical example. "Ao Dai Shop A" has three shops in Ho Chi Minh City, a Facebook page and a stall on an online marketplace. All four sell one line of ao dai, each style in five sizes. The owner wants to make four promises to customers. For each, the question is which stock data is needed.

Each row asks for more stock data than the first: counts per shop, then holding stock, then sight of other shops' stock, then a process for booking returns back in. The Inditex report ties RFID to tracking each garment and to its integrated stock system, then lists services similar to these four promises (Inditex, 2019). I think a small shop cannot easily do the same with RFID tags, but the table shows which promises can be kept with less data.

For three shops and one online stall, I would make only the first two promises, and only if sales at all four places go into one spreadsheet or system and are deducted at once. If updates happen only once a day, after closing, I would promise nothing beyond "call before you come". That is my judgement from the order in which the table's data demands rise, not a measured result. It suits a shop with few sizes and sales that do not arrive in bursts. It does not suit one running deep, time-limited discounts, where many people chase the last item within minutes.

Going deeper: what I read and what I did not

I read the full introduction to the special issue by Verhoef, Kannan and Inman (Journal of Retailing, 2015, volume 91, issue 2, pages 174 to 181) from a public PDF on the lead author's university repository, including the definition, Table 1 and the summaries of the studies. I have not read the original papers by Cao and Li, Herhausen and colleagues, or Pauwels and Neslin, so I report only what the introduction says about them. The five-step path in the figure is the library's illustration; I did not find it in the part I read.

I read Inditex's 2018 annual report through text extracted from the company's PDF, reading closely the passages on RFID, integrated stock, Click&Collect, delivery and online sales, not every page of this long report. Two cautions. First, it is a company report, so every judgement of effectiveness is self-assessment. Second, it gives the online share on two bases (12 per cent of net sales and 14 per cent in markets with online sales), so I give both.

I could not check the financial value of an integrated stock system, because no source in the library measures it. The "Ao Dai Shop A" example is hypothetical, to show how to frame the question, not a tested recommendation.