By Lawrence Garvin, Head Geek, SolarWinds
With demand for online deals and sales events growing in Australia, catering for peak demand online is increasingly on the minds of many retailers. However, it’s important to realise that the fundamentals of the e-commerce store are not so different from its bricks-and-mortar predecessors. The major differences are in scale and speed of responsiveness to demand: an e-commerce shopfront can scale up to meet peaks far faster, and to far higher volumes, than its physical counterparts. Moreover, e-commerce stores face extremely high expectations from their customers when it comes to 24×7 operations. Retailers must support their e-commerce platforms with monitoring tools that track demand; as well as cloud services that can ramp up to meet excess demand when those monitoring tools tell them to.
Sales forecasting for the retail cloud
Sales forecasting has always been a critical component of the retail industry. If a retailer can predict sales in peak periods like Mothers’ Day more or less accurately, it can ensure the necessary resources – inventory, staff levels, supply chain capacity – are in place to maintain business-as-usual services standards. However, this traditional approach only works for “known unknowns” like public holidays or other recurring events; for less predictable spikes in demand, retailers need systems which can detect growth early and reprovision on the fly.
E-commerce sites can do this far more effectively than bricks-and-mortar stores: you can’t, after all, add another aisle or extra floor space by the time customers are streaming through your door, but you can do the online equivalent in seconds. However, that’s exactly why consumers have such steep expectations from their online shopping experiences: seamless transactions and customer service at any and every time are expected as the norm, not the exception. A failure to deliver can result in widespread loss of customers and credibility.
Retailers need to install software and systems that monitor real-time demand across their e-commerce platforms. The best monitoring tools will not only track traffic metrics for websites and payment channels, but also automatically and identify respond appropriately to different scenarios that e-tailers might face. Tools like SolarWinds’ Web Performance Monitor, for example, are able to differentiate between peak retail demand and other causes of high network traffic – particularly malicious attacks like DDoS attempts – based on a number of metrics including actual sales volumes and completed transactions, rather than just page visits and shopping card additions.
Network and website monitoring tools should supplement traditional sales forecasting. Retailers need to first assess what constitutes “peak demand” based on their customer base and IT resources, before programming these thresholds into their monitoring tools. They also need to consider how effectively these tools can communicate with other critical elements of their e-commerce infrastructure, particularly the cloud hosting which ultimately determines their ability to meet demand.
What’s the chance of a crash?
With the majority of e-commerce store hosted in the cloud, retailers need to ensure storage, bandwidth, and processing resources can all rapidly scale up at the first sign of peak demand. Best-practice IT provisioning usually ensures a buffer between current loads and total available capacity at any given time, so that the site never gets close to exceeding allocated resources. That should be a priority in any conversations between retailers and their IT providers.
What about when the retailer isn’t in control of the e-commerce platform? Retailers should still apply due diligence to the online event organiser’s actual capacity and the risk that it may be oversold; that due diligence should then inform how the retailer negotiates and assesses SLAs, contracts and other guarantees of platform performance. Doing so will help retailers avoid online sales events which may crash as a result of oversold capacity (the problem here is different from a platform’s inability to scale, since levels of demand could have been easily predicted).
Retailer need to apply the same rigour of sales forecasting to their e-commerce platforms that they already do for bricks-and-mortar stores. Doing so will inform not only how they use monitoring tools and cloud services to flexibly adjust to peak demand, but also assess the reliability of online sales events. The goal of any e-commerce platform is ultimately to elevate customer service and satisfaction – not to go out with a bang.