Enter a search term such as “mobile analytics” or browse our content using the filters above.
That’s not only a poor Scrabble score but we also couldn’t find any results matching
Check your spelling or try broadening your search.
Sorry about this, there is a problem with our search at the moment.
Please try again later.
When online photo company Alamy decided to expand its business globally one of its most important challenges was maintaining a decent site speed to ensure a consistent user experience for global customers.
As it operates in a crowded market place, Alamy focuses on differentiating itself from the competition both on the size of its database and the level of service it provides.
An important part of that is the overall speed of its site, the homepage and an on-site search tool.
So when Alamy found that its US site was noticeably slower than the UK version it ran a series of tests using Compuware to work out the best way of getting its US customer base up to speed.
To find out how Alamy resolved the issue and how it continues to track its site speed I spoke to head of production Richard Taylor...
What did the analysis of your site reveal and how did you improve the performance of the US site?
We initially started working with Compuware Gomez to help us understand the performance of our US website and assess various options for investment to get it to match the speed of the UK version.
This involved looking at several solutions such as building data centres and working with suppliers who could distribute content on our behalf.
The analysis showed us most of our customers are in New York or LA, so to achieve the best results we needed to host our content in those areas.
We decided the best option was to open new data centres on each coast, and as a result the homepage in the US now matches our UK target of two seconds.
Within the on-site search results we set a target of 0.2 seconds for the time it takes an image to enlarge when you hover the mouse over it.
Prior to opening the data centres the ‘zoom time’ in New York was 0.7 seconds, while our competitors were anywhere between 0.3 and 0.7 seconds.
Once we got the equipment in New York our zoom time decreased to 0.1 seconds, so that’s the difference of having your data hosted over there.
How slow was the site prior to opening the new data centres?
Although the load time was similar to the bulk of our competitors in New York, we were in the lower quartile.
Our customers tend to be very busy people working for newspapers and other media outlets, so any slowness is a problem for them.
Have you since identified any other areas of the site that you needed to improve?
One of the downsides in spreading your infrastructure is that it becomes more complex, and sometimes things don’t quite go as you plan.
Data from one our of LA nodes (analytics tool that sits in the data centre) showed that our users in LA weren’t actually pulling data from the local data centre.
We sent a team out to fix it and they reported that they’d brought the data centre back online, however in actual fact the problem hadn’t been rectified.
The only way we knew that was by having the external monitors in place with Gomez.
Also, sometimes we have customers that telling us that our site is slow in one area of the world, and we can use data from Gomez to tell them that it’s actually an issue with their ISP.
You recently expanded into Australia as well. How do you ensure that you achieve a decent site speed over there?
In the UK our site speed is one second or less, but in Australia it’s around three seconds.
Content on the Australian site is pulled from the LA data centre, so in order to speed it up we are trying to reduce the amount of data required by lowering number of objects on the site and size of them.
How does the slower site speed in Australia impact conversions?
We can’t measure the impact on conversions at the moment, but studies do show that the slower speed is likely to mean we lose a certain number of sales.
However when we speak to sales guys over there they’re not bothered about the performance of our site as it’s similar to our main competitors.
But in the UK we’re currently achieving sales growth of 30% per year, and to achieve similar numbers in Australia we know we need to improve the site speed.
How often do you check the site speed, and what action can you take if you think it’s maybe too slow?
We have a number of monitoring systems in place so we’re constantly measuring the site speed.
Gomez is our ‘out in the field’ measuring system, but it’s probably about once a quarter that it tells us something we didn’t know about.
We use Google Analytics as well; they’ve recently introduced a system that tells you about the speed of pages which is very useful.
Do you measure site speed on mobile? If so, what tools do you use?
We don’t track mobile speed at the moment, but we will probably have to start doing it as our customers’ needs change.
Our users falls into two categories - customers and contributors – and our mobile use is principally from contributors.
I think that’s because a majority of customers are home or office based, so they are picture researchers working on a desktop.
Our contributors are photographers so they tend to be out and about more and use mobile devices to track how well their images are selling.
Over the past 18 months our mobile traffic has gone from 1.5% to 4% of overall site traffic, and we know we need to do more work to support those modes of use.