Workshop: A beginner’s guide to online disasters — how to avoid them.
There’s plenty of talk about how to achieve fantastic things on the web, but what if things don’t exactly go according to plan? Accidents happen. Jonathan Crossfield looks at how to avoid internet disasters.
The British novelist G.B. Stern wrote, “The optimist invents the airplane and the pessimist the parachute.” But is it pessimism that motivates you to be prepared for disaster, or mere practicality?
With all this talk of how easy it is to be highly successful online, and with so many tools available to monitor and analyse every trend, it is possible to be tricked into viewing the best case scenario as a foregone conclusion.
Where’s the parachute?
Planning your online strategy based purely around best case scenarios can expose your business to considerable risk. In the battle between budget and infrastructure, it often becomes harder to justify paying for what seems unnecessary when everything is going smoothly. But this kind of gambling – although extremely common – can prove far more costly over time.
You might jetset around the country every week for your entire working life and never experience so much as a stale vol-au-vent, never mind a plane crash. But would you therefore dispense with the life rafts, inflatable jackets and masks falling from the ceiling in return for a cheaper fare? Of course not. We would never take such risks with our personal safety, regardless of the low statistical likelihood of disaster.
We understand that a positive personal history of successful flights doesn’t reduce the risk of something unexpectedly going wrong next time you stow your table and strap yourself into your seat to be hurled at brain-melting speeds through the sky inside a large metal dart.
Business disasters might not be quite as serious as a sudden high speed stop on the side of a mountain, but then the remedies are also far easier and cheaper. So why risk your business by not planning for how you could recover from the worst, even when things are going superbly well? Where’s your parachute?
“Where did my website go?”
A crash of a different kind can be a disaster for any online business.
As online business continues to evolve and new methods of attracting attention are developed, website traffic has become far more volatile. Traffic can run steady for days at a time, until the day a link appears in just the right place for thousands of people to jump on board at once. Social media has increased this possibility, where a single tweet on Twitter, vote on Digg or mention in a popular blog can start a chain reaction causing a massive burst in traffic.
Earlier this year, popular T-shirt site Threadless announced a one day only sale – every T-shirt was reduced to only $9! Threadless is extremely good at self-promotion and effectively used email marketing and social media to build buzz. The $9 sale message spread very fast and very far, resulting in a huge spike in traffic to the Threadless website – including a lot of eager first time visitors, credit card in hand.
When the traffic spiked, it should have been time for the Threadless team to pop the champagne corks for a marketing job well done, but it quickly turned into a disaster. All that simultaneous activity placed incredible strain on the Threadless servers. Their hosting service couldn’t cope with all the transactions trying to squeeze through, so the servers crashed. And once the servers crashed, there wasn’t any selling of $9 T-shirts going on.
The servers are the hard drives where your website information is stored. Every time a person interacts with the website, it sends a request for information and then returns that data to be displayed in the browser. Take that server offline and your online business disappears until it can be restored!
Bandwidth is the ‘pipe’ that connects your website on the server to the internet browser on your customer’s monitor. It is a measure of the amount of digital information that can be accessed in any one moment. For example, if your main webpage is 300 kilobytes in size, then every time a web browser accesses it, 300kb of data travels down the ‘pipe’. How fast this data is transferred depends on how many people are accessing data down the pipe at the same time and how large your pipe is.
If your ‘pipe’ can cope with a certain amount of data transfer per second, but the number of people trying to access your website exceeds this data limit, problems begin to occur.
So it was with Threadless.
Even after the server was back online, it was incredibly slow to load and process customer transactions. For most of the 24 hours of the Threadless sale, accessing and buying a T-shirt from the site was a painful, if not completely futile, activity. Threadless’ brand image was hurt. How many of those customers who never got to buy a T-shirt, as well as those who did but were frustrated by the experience, would be discouraged from returning in the future?
Many hosting companies store websites together on a single shared server. This means that if one website experiences a heavy load, it can crash the server and affect all the other sites stored with it. By comparison, large hosts such as Netregistry store websites in a clustered hosting environment, preventing the performance of one site from affecting those around it.
Different hosting environments have alternative configurations. It is wise to check with your provider what to expect should you experience a flood of traffic. How many connections to the internet (or pipes) are available? What redundancies are in place to reduce the risk of crashes?
Threadless was wise enough to email its customer base the following day to apologise and extend the sale a few extra hours. But apologising to customers is not a symbol of a successful online strategy. With a bit more planning to ensure the servers and bandwidth could cope with unexpected spikes in traffic, Threadless could not only have avoided a disaster but raked in all those extra sales.
“My bill is how much?”
Crashing the server is not the only risk when chewing through data in a traffic spike. If your hosting account has a monthly limit on data transfer, unexpected spikes in traffic that take you over that limit can cause headaches.
Similar to your home internet or mobile phone accounts, you have a capped limit on your hosting service. Go over that capped amount and two things can happen:
- Slow data transfer – some web hosting providers ‘throttle’ the data transfer speed down to a crawl. This means your website will become slow to load for customers – go-and-make-a-cup-of-tea slow. Most customers won’t stick around.
- Additional fees – other providers, including Netregistry, don’t throttle your bandwidth. Instead, they keep it going, to keep your website operating at maximum efficiency, though you can be charged extra for data transfers above your subscribed limit. Receiving an unexpected larger bill at the end of the month for every extra MB transferred could seriously undermine the success of your promotion.
Netregistry also allows you to buy extra data traffic allowance in advance at a discount, if you think a possible spike could smash your capped limits.
“Help! I need somebody!”
When the worst happens, who will you call? Imagine a nightmare business day. Your website might be down without explanation or someone may have hacked into your system. Maybe a third party application has created security issues with your payment gateway or your email is no longer getting through.
When the panic sets in and the realisation that every minute that passes is causing inconvenience to your customers, how much is a decent and rapid support service worth to you? How much do you stand to lose if your website is down for a few hours, a day, a weekend, longer?
Yet, when most people choose their online services, they don’t consider their nightmare business day. They’re thinking of the wonderful possibilities and how smoothly everything will work. All of those possible horror scenarios seem distant and unlikely – much as an airline crash is not something most of us think about when boarding a plane.
So, when deciding on the right package and supplier, support considerations can often be overlooked or ditched in favour of a cheaper deal.
Have you ever thought that email-only support will be sufficient? Surely you don’t need 24 hour, seven day care? Surely it won’t matter that the support centre is overseas in a wildly different time zone. After all, you don’t have a problem to be fixed. What could possibly go wrong?
And then it happens. Welcome to your nightmare business day. Suddenly, you’re caught in a world of digital pain, made worse by talk of code and jargon and weird technical issues that may as well be another language.
Calling the other side of the world to try and fix the problem isn’t much fun afterall – especially if the differing time zones mean that your provider is possibly asleep. Seeking help by email doesn’t ease the blood pressure either, waiting anxiously for a response as the clock ticks and the customers complain.
Choosing a local provider has obvious benefits beyond just ease of contact in your timezone. For example, a US hosting company may schedule necessary maintenance – requiring server downtime – during their quietest periods. But, that may be your busiest period here in Australia.
Also, when disaster strikes, timely responses are crucial. Waiting for emails to be answered – especially when you are contacting a provider out of business hours, can extend the disaster for hours. There is no substitute for being able to call up a local support member 24 hours a day, seven days a week who can quickly put their finger on the problem.
You may only require this kind of disaster assistance once in the lifetime of your business, or you may require it frequently, depending on the unique circumstances of your situation – but even one mishap prevented or rapidly remedied could save your business from serious financial loss and damage to your reputation.
“Who put that code there?”
Even if your website stays online, active and turning over a steady flow of traffic, you might not be safe. Every day, websites of all kinds have their code hacked by unauthorised individuals keen to steal data or customers.
The results can be devastating to a business. Should your website be compromised, extremely serious breaches can include theft of customer account information and credit card details. Such attacks can destroy your business reputation and can lead to criminal charges if you have not taken sufficient precautions to secure your site and protect private information.
Sometimes a hacker may not attempt to access data but simply place additional code into your site. This can commonly be a piece of advertising script, earning money for the hacker every time one of your customers clicks on it – until you notice. Or it can simply be vandalism – changing your home page or destroying your layout for whatever malicious reason the vandal may have.
Common website hacks involve exploiting holes in popular blog or third party website creation software. The developers frequently release new versions to plug these holes and thwart the hackers, but if you fail to update as new versions become available, you could be leaving a vulnerable back door open into your site.
Hackers are highly ingenious and skilled at finding new vulnerabilities to gain unauthorised access to websites – that’s part of the thrill of the hack for
them. Regularly change your FTP passwords (FTP is the File Transfer Protocol; it’s the means of transferring files from your computer to the website server). Always ensure the passwords cannot be guessed or easily determined. Also, regularly ‘spring clean’ the site by archiving redundant web pages, cleaning out superfluous code and ensuring everything is running at tip top efficiency. Fewer bits of redundant code mean fewer vulnerabilities.
“Phew – that was a close one!”
It is always advisable to regularly perform secure back ups of your website data in a second location. If you are unable to, arrange for your web designer or the person responsible for administering your site to organise a regular backup schedule. Should anything happen to your hosting server, having your site backed up on a separate computer means you can quickly restore it, repair lost data or create mirror versions of the site on other servers to get things moving quicker.
No one wants the worst to happen and we always hope that it never will. But failing to plan for the possibility that something could go wrong not only
puts sales at risk, but may also risk the loss of your customers, your reputation, your finances and your entire business operation.
Have you prepared your parachute?
How much uptime is good enough?
The last thing your website needs is downtime, which is why many hosting providers offer uptime guarantees. But how much is good enough?
The difference between 99.5% and 99.9% uptime might not seem very much, but actually equates to 42.9 hours of your website being unavailable to customers each year. That’s nearly two days of forced closure with no sales and no money coming in. Therefore, 0.4% of a difference can cost your business a lot more than you may think.
Jargon translator
- Web host: The company that administers the server where the webpage is stored.
- Server: The specially configured hard drive that stores the website pages and serves them up to internet browsers when the appropriate address is typed in. A ‘shared server’ stores many different hosting accounts. A dedicated server is assigned to a single large hosting account.
- Bandwidth: The amount of information that can be accessed from the server at any one time.
- FTP: File Transfer Protocol. The method of transferring files and data from a computer to the server and visa versa. This is a password locked, secure connection
- Short-tail phrase: A key phrase of only one or two words, resulting in a much wider and longer list of results as well as greater competition from other websites.
- Natural results: The results generated by the specific keyword request.
- Sponsored results: Paid advertisements that appear next to the natural results and are triggered by the particular keyword or phrase used.
Build your survival kit
Firstly you will need to ensure you have your online backups. A cloud solution is ideal in this situation.
Using an external hard drive is still common practice and still very reliable to work in conjunction with your cloud backups.
Have a backup company on standby. If you or your team are unable to be available, having a contracted service provider available will reduce a lot of stress when things go wrong and you may have a wider set of skills available.
Speak to our team if you are interested in website maintenance, server maintenance or learning more about backup solutions.
