In this series I’ll take a look at cloud and large data; what it can offer retail to optimize costs and how to facilitate a customer centric strategy.

Riddle 1: How to scale or shrink on demand?

We have probably all spent agonizingly long hours in traffic jams sometime or another. Wouldn’t it be great if the road would suddenly become wider until the blockage is dissolved and then shrink back to its former size? I would then get home quickly and of course adjourning natural habitat or homes would not get destroyed. Unfortunately this is not possible in the physical world.

The virtual world, however, is capable of providing exactly this feature – scale on demand when there is too much traffic and shrink when there is only passing trade.

Planning for peaks is for the birds

A few years back, when planning IT systems, we always had to plan for peaks. A peak by definition is a very narrow time frame, thus in the remaining time there was always an excess of hardware, costing our clients tens of thousands of dollars for overhead costs, just for these few important moments in time. Now with cloud technology available, these days are over. As with electricity you now pay for what you actually use and you don’t have any shortage of supply. Do you need your house fully lit up – go ahead and switch on all the lights. There won’t be any shortage, as long as you pay your bill. You, however, are in control. The moment you switch them all off again, electricity usage goes down instantly without you paying for any additional power the plant might be producing.

For companies such as a vente-privee, who do flash sales, this is a blessing. Since stock and time of an offer are strictly limited, the moment vente-privee sends out an email huge numbers of customers swarm onto the servers. The goal is to get customers to place the order; that’s all that matters. No browsing or shopping around. And that’s where vente-privee makes it’s money. So every customer who comes to the site but is rejected by the overloaded system is a lost sale. Whereas on ticket sites you can put customers in a cue and let them wait, this is not appreciated by these customers as they may just be missing the deal of their lives. Cast your mind back to the news reports of the Boxing Day sales, picture that online if you dare.

Season’s greetings

Though it’s not just for these types of companies, many retailers make their money dependent on the season. For most Christmas is the busy time and that is when they need more power. Now Christmas for retailers runs much longer then for us the customers – for many retailers it starts in September and then abruptly ends on December 24th. However, that’s still only 4 out of 12 months when additional hardware is needed. Thus with the old approach a retailers operational costs was much higher in the other 8 months then they ought to be. With cloud technology seasonal demand can be smoothed out almost like flipping on or off a switch. I still remember how difficult it was adding 6 more servers during the Toys’R’Us launch in 2007, because traffic estimates were way to low. Just ordering the boxes and getting them into the data center took a few days. (Luckily the hosting provider had some boxes in some storage we could use immediately. Otherwise it would have taken even more time to order them). Then we spent 48 hours to set them all up and get them live, this process was painful for everyone involved. It’s slow, inefficient and very costly (short and long term). Adding those 6 boxes today would be a matter of a few clicks and 10 minutes or so waiting for them to boot up. That’s it. And when they aren’t needed anymore, they can easily be powered down in the same fashion.

To infinity & beyond

Storage is another challenge. Pictures, videos and other rich data add up quickly. Although much is stored on streaming servers many media files will still end up on your system. And when you encourage your customers to create and upload their own content data, volume gets quickly out of hand. So scaling of your disk capacity becomes a challenge and the cloud helps here too as storage size may be of the cloud operators concern, not yours.

The next article in this series will be about how to follow a customer journey across all touch points and how big data can help achieving this.