If everyone working on a system is trying to stabilize and improve the rate at which tested, integrated software is successfully released to end customers, then teams across the system will naturally have to work together and will find that their goals are compatible.
Tags: lean TOC scale

Ben Christensen - Reactive Service Levels. Fantastic talk on dealing with failure - Netflix style.

The true purpose of sales is to create new value for customers. Especially in a startup or growth company that’s addressing a new market or trying to solve a complex problem. That’s why enterprise/SaaS sales requires a well-developed process and guidelines.
— If SaaS Products Sell Themselves, Why Do We Need Sales? - http://a16z.com/2014/05/30/selling-saas-products-dont-sell-themselves/

Open source technologies are the future for online gaming

(note: this article was published in the March/April 2014 edition of iGaming Business magazine)

In the almost twenty years since the first online sports bet was struck, operators have both looked forward to and feared seasonal spikes in activity.

Major events like the Cheltenham Festival, Grand National and major football tournaments offer unparalleled opportunities to boost profits.

Yet these random peak demand events also put considerable strain on a website’s infrastructure and have led to some spectacular crashes in recent times, particularly when accompanied by punter-friendly marketing offers.

This summer’s World Cup in Brazil is expected to be the daddy of them all when it comes to customer acquisition and retention – but you can bet your boots more than one IT department is praying they can cope.

Part of the problem with peak capacity requirements, of course, is gauging just how serious demand will be.

Purchasing hardware and rack space requires a good deal of forward-planning, not to mention considerable expense. Despite this hassle, there remains the dilemma over whether a sportsbook has acquired enough capacity or, indeed, has wasted money by panic buying too much.

The good news, however, is that there now exists a far more reliable, cost-effective alternative, which ought to make serious outages a thing of the past.

Open source technology, which is widely used in other technology-driven industries, offers the proven ability to handle these peak-scaling requirements without the headache.

Such technologies as OpenStack and MongoDB, which we use at Sporting Solutions, allow horizontal scaling, which can be automated utilising trading history around events to provide metrics on how much demand was actually put on the trading systems.

That data can be used to build predictive algorithms that automatically scale systems up or down based on expected and actual demand.

This takes the manual element out of the equation and ensures that the system does the work itself, removing in one fell swoop the issue of hardware planning and offering optimum efficiency. For the CFO’s reading, using private cloud capacity, such as OpenStack, can be far more cost-effective than using public cloud when running systems 24 x 7. That’s not to say public cloud doesn’t have its place and is useful for bursting requirements, which can also be part of the automated scaling process.

As someone who has worked for technology companies outside betting and gaming, it is surprising that few if any of the leading sportsbooks have embraced open source solutions thus far.

Some are further down the road than others, but it is hard to pinpoint any who are not wedded to throwing hardware at the problem via vertical scaling.

In fairness, many are saddled with legacy systems that their predecessors have bolted on extra capacity to when they’ve begun to creak or, worse still, fallen over.

Others have been encouraged to see it as a cost centre issue, or one that should automatically be outsourced rather than dealt with in-house.

I’d argue, however, that if they have ambition to innovate and be competitive in the industry, when it comes to the performance and capacity of their site, they need to start thinking differently.

In fact, it would do nobody in our industry any harm to take a leaf out the books of the likes of Google, Facebook, Twitter and Amazon and see how open source is at the centre of their technology strategy.

These Silicon Valley heavyweights continue to secure a large proportion of their revenues by operating at massive scale and seamlessly handling peak demand very much like we see in our industry.

The technological know-how embedded in the workforces of these companies has obviously helped them spot the competitive advantage available.

But the same option is open to sportsbook operators if they grasp the nettle and are prepared to be innovative in their approach.

Technology has taken a back seat for too long in the gaming industry – apart from the select few like bet365 who have invested heavily in it via proprietary software, and reaped the benefits.

Open source software allows operators to benefit from the faster innovation cycles that are offered which can be converted into business value, not to mention the benefits that come from being associated with the software in terms of recruitment and brand recognition.

In addition, they will no longer be reliant on vendors telling them how much money they have to spend on extra licensing, or beholden to an out-sourced solution charging a premium for its services.

Unless the present position changes, sportsbooks will continue to crash during peak times, costing money in lost revenue and leaving them looking for a quick fix to plug the gap.

As a result, customers will continue to experience a poor service and be frustrated by their inability to take up an offer, place a bet, or collect their winnings.

If operators are lucky, or have deep pockets to buy traditional hardware and proprietary software, they might yet get away with it for a while longer.

But with the World Cup looming large on the horizon, keeping their fingers crossed doesn’t feel like the most sensible option.