Where are we? A brief overview
In more mature industries we are privileged (cursed?) to have a common concept of a coverage check, or ping. This simple operation is instrumental in the trading of leads amongst friendly competitors in a given industry. Higher customer expectations and rising costs for organic consumer aggregation dictate a growing need for this additional facilitation on both sides of the lead. However, one of the many problems everyone will encounter on a regular basis is timeouts or expired communications between 2 or more parties. In the following sections we'll discover some of the reasons behind timeouts, their implications, and how to counteract and avoid them.
The two timeouts
There truly are two forms of timeouts that you should be aware of, timeouts that you invoke on your buyer network and the timeout each provider in your provider network imposes on you.
Let's start with buyer network timeouts and stating a hard truth, your buyer network is slow, really slow. There's no way around it, 3rd party coverage checks will always be inconsistent and slow so long as you are connecting to external software, hardware configurations and business rule sets. You can't, and probably wouldn't want to, modify or manage your buyer network's systems, so it is always best to treat them as suspect and keep a tight grip on all aspect of your relationship including time outs.
Provider timeouts are where the problem really resides. When you time out to your provider you essentially lost a sale, this is how it should be treated. It is essential to know as much about your provider as possible and to always give a response in every scenario. One consistent truth about most industries is that as they mature everyone works with everyone, and many times a single poorly performing system creates a domino effect throughout an entire industry. In scenarios such as this, it is common to see large fluctuations in traffic (you may see 10x the traffic from one provider for one day, and one day only), and many times it is the tightly woven systems that benefit and the open ended systems that lose out.
Tips to increasing throughput by limiting timeouts to your provider network
1. Know your providers! Knowing your provider is instrumental in configuring a relationship. What is the timeout they'll impose on you (they'll tell you this), do they share many of the same providers and buyers that you currently have (they WON'T tell you this, but little industry knowledge and good insight into your ping metrics reveal this easily), and the volume of traffic they will be sending (again, insight required) will give you the information you need to eliminate nasty problems like redundant and circular pinging.
2. Support provider based timeouts to your buyer network. If you know your provider will time you out after 5 seconds, impose a 4 second timeout on your buyer base for that relationship.
3. Know the throughput capabilities of your own software; don't just throw money at hardware as a solution. This is by far the hardest to measure. Effective thread management is extremely difficult to master, most developers will tell you what they have developed is gold, but the truth of the matter is almost all homebrewed threading solutions crumble under pressure. Additional hardware (load balancing) helps, but is expensive and a band aid solution. More on this in a later article.
4. Experiment. Try reducing your default timeouts, try increasing them, and find the sweet spot. Keep in mind, this may change as time passes, what worked a year ago, probably doesn't work now!
5. Reports, insight, analysis. Almost a duplicate of #1, but good insight into your system is absolutely crucial, I cannot stress this enough. If you don't know your average ping times from your buyers and to your providers you're likely already having throughput issues.
Timeouts are here to stay, for good
At the end of the day, you can only do so much; your provider base will still time you out. But what is important is that you do your best to stay off the dead beat list of companies in your industry who are constant headaches to everyone. Every industry has a hand full and working with these companies is always a double edged sword. The best thing you can do is stay on top of your own system and stay agile in the pursuit of faster responses from your own system through the careful handling of all bottlenecks in your system.