1. We'll ALL be talking about BizOps
The massive digital transformation we are in the middle of will compel business and technology teams to work closer together than ever before, and data scientists are the social glue who will make it happen. They will do it through the relentless discovery of strong correlations between business metrics and technical metrics. By the end of 2017, this new culture will finally be mainstream (although everyone will be playing catch-up with Amazon for years to come) and we’ll all be talking about “BizOps”.
2. Retailers will realize they’re all competing with Amazon
Ridiculously scary, but true. Amazon has moved into a staggeringly broad range of markets – including music, cars, clothing, photo processing, even lumber. No matter what kind of business you’re in, it’s safe to say that Amazon has its eye on it. And while people may love your brand, they’re not at all loyal to your website. If they can buy your products through Amazon Marketplace, they will. That’s because they trust that Amazon will give them a safe, fast, reliable customer experience. Amazon has set the UX bar extremely high, and shoppers today expect every retail site to reach that bar.
3. The competition for holiday dollars will continue to move online
This past holiday season, Black Friday traffic exceeded Cyber Monday traffic by 21%. Most shoppers are tired of the crush of malls, and these increasingly savvy consumers know that in-store deals are often no better than online deals.
In-store shopping on Black Friday – and over the holidays in general – will continue to decline, hastened by the prevalence of mobile devices and ubiquitous free shipping. Retailers need to be ready to meet the rush. Shoppers expect sites to be always available, even during record-setting traffic peaks. And those record-setting peaks are going to occur even more frequently in the future.
4. We’ll manage performance like we manage products
Prioritisation will become key. As consumer expectations for speed increases, engineering and DevOps teams will need to manage performance like a product. Using data from monitoring products and advanced analytics to shift from fire fighting the most recent issues or optimising the slowest pages to identifying the key areas that need attention. Focus for development, testing and optimisation, will be on the pages, use cases and demographics that create the most revenue.
5. Data integration using data API(s) will become more common and will happen in near realtime
Traditional ETL and data warehouses have a lot of friction and entropy. Stream processing and data integration in the stream will peek its head into the market as a preview of powerful capabilities in the future. (Big Beacon and BlockChain are both examples of this.)
6. Machine learning will move out of the lab and into the real world
Machine learning models will be trained and installed in production systems to predict outcomes and to detect abnormal events and system states. This will result in faster and more targeted responses to systems issues. Problem alerts will carry context created from Big Data to support quick diagnosis and remediation. Advanced alerting and web integration (webhooks) will tie all of this together to “machine assist” the diagnostic and remediation activity.