On the web: Towards a global stream

With employees sat at around 13,000 desktops all over the world, Reuters needed to find a way to reach them all to communicate important company messages in an efficient and direct way. Steve Clarke explains how, after some initial teething problems, the company developed a live streaming and video service that now represents a powerful means of communication.

After 20 years in broadcast television, if there’s one thing I know it’s that everyone is an expert on TV. From the pub to the supermarket queue it’s picked apart, criticised, mimicked and mocked. But only those who’ve worked in it know how hard it is to get a quality image on a screen, transmitting content that someone actually wants to watch.

Of course the advice of others is kindly meant, and when I took over a new role as head of internal media at Reuters three years ago many who knew me in the company (and knew that I’d once worked in the television industry) assumed that I would be the man to put wall-to-wall internal TV up on our staff intranet within a matter of days. I probably assumed it too, but, as they say, nothing’s ever quite that simple. What lay ahead was a long road, but one that has left Reuters with 80 per cent of its global staff connected to desktop video available both on demand and, more recently, live.

There was no doubt that Reuters was a prime candidate for the installation of some sort of robust video system at this time. For a truly global organisation with, at that time, some 18,000 staff spread across 150 countries – and with most of the staff facing PCs for at least some of the day – video was a potentially accessible and unifying communications tool. It was also clear that existing attempts at video communication weren’t working.

Earlier attempts had involved making a video in London, copying it onto hundreds of tapes in multiple formats and mailing it out to offices around the world. Were they watched? I, for one, doubted it. Many offices didn’t even have a tape player – why should they? – let alone the time to unwrap a parcelled tape from London, gather round the screen and view it like a stereotypical 1950s family.

The answer was to make the tapes available on staff PCs so that short, simple video messages could be just a click away wherever Reuters staff were. The question was, how to do it? As a former TV reporter I wasn’t too worried about getting the content together but the knotty problem of how to get it to the viewer was a bigger conundrum. After all, budding TV producers aren’t normally asked to build the transmitter as well as make the programme. But for those who enter the world of corporate internal TV, beware: every problem will be yours to solve, from ‘soup to nuts’.

Undaunted (due to complete ignorance) I thought I saw an easy way to solve the video problem at a stroke in early 2000. We took a piece of video content, encoded it and placed it on the main Reuters internal staff site – which I was responsible for – so that our employees could download it at a click. Disaster! As our staff rushed to try this new wonder, enthusiastically demanding the fat video files on their desktops, megabytes of unexpected loading was placed on our intranet. Not only did this mean the video didn’t work properly for many – it also threatened the integrity of the whole system. Well, I was new to the job.

Memos from some of Reuters’s most senior global webmasters quickly pointed out my error and helpfully offered some rapid personalised retraining methods. I feared that my wife would have had trouble recognising me without the aid of dental records once they had finished with me.

However, they turned out to be a friendly bunch deep down and once I got to know our senior technical staff they gave me some very good advice. A solution needed to be robust, truly global and would require additional technology to supplement the servers that already made up our intranet. Installing it would take time and money. They could provide some of the time, in terms of effort, but the money would have to come from me – apart from that everyone thought that the idea of video for all staff was a great one.

In a sense we were back to square one, but this time we had technical buy-in and a real feel for the size of the job ahead. Reuters chief technology office even responded with some seed money to get an evaluation project going. This project tested the equipment and software offerings from outside suppliers that could support and manage a global video system. An official internal project – known as project Lightpath – was set up. I was the client and was asked to draw up the specification as to what its key qualities would be.

To this day it was my firm belief that this is one of the major reasons that project Lightpath has been relatively successful. Technology is often created or bought in without a clear idea of what the client or end-user really wants. By specifying the exact requirements we ended up with technology that really gave us what we needed and fitted in well with existing technology already in place at Reuters.

What I specified in early 2001 was the following:

A good-sized picture (not an internet style postage stamp);
A steady image (no flashing or sudden freezing);
Reasonable picture quality (we settled on streaming at 250KB in the end);
An ability to retrieve the material globally at consistent quality;
No requirement for live streaming.

The last condition may seem a strange one, but at that time, shaping the system to stream live looked as if it would compromise almost all the previous requirements. Also, as we were just dipping our toes into this medium, live streaming was not high on my list of priorities. I was quite happy to cache pre-recorded material. Later things would change – but I’ll come to that in due course.

By January 2000 we had a project running and a tentative plan (at least in my own mind) that we would launch some time in the summer. The only problem was that I didn’t have a budget. It was around this time that I had a meeting with our CEO, Tom Glocer, on an unrelated subject. I took the opportunity of asking if he would like to appear before staff via video on the day of our results announcement in July. “Sounds interesting,” said Glocer, a self-confessed lover of technology. It wasn’t a budget but it was something even better: buy in from the top.

To cut a long story short Glocer’s support allowed me to raise a budget, albeit not a big one. I’m not going to give an exact figure here but I can say our video installation has been a low-cost enterprise from the start and I’m pleased to say it’s staying that way.

Armed with some hard cash I went back to our technical staff to see what kind of solution they were proposing to the criteria they had already set out. The answer, in a word, was Cisco. The company was already a strategic partner of Reuters and its technology fitted in well with solutions we already had in place. It was also felt, by our technicians at least, that Cisco had a system that was most likely to be capable of evolving over time, a belief that proved to be true in due course as you’ll read later.

A short description of the specifications of the Cisco system is a ‘cloud’ of caching servers linked from London via key Reuters hubs like Geneva, New York and Singapore to smaller servers – essentially one per office location. At the heart of it was a Cisco content-distribution manager sitting in one of our Docklands offices in London. This judges when and how to distribute material into the system so that high priority traffic is not disrupted. Via this system material would be sent to the caches globally in offices like, for example, Tokyo, from where local staff would pull it onto their screens at high quality and on demand. Feeding it all was an encoder connected to a tape player and storage system in our head office in Fleet Street. The set-up has proved simple but effective to this day.

To return to the story, in March 2001we agreed an aggressive testing and installation schedule that would see the whole set-up working in time for our first broadcast tape – a video message to staff from our CEO – on the day of our results announcement at the end of July. In retrospect setting a ‘must-do’ date by which the system had to be fully working, with little fall back if we failed, was rather shredding for everyone’s nerves, but it did concentrate minds.

As a result our global network of technical staff functioned perfectly, the system was fully installed and tested for the first time the day before the results announcement. That night we downloaded the video message from our CEO in London and waited with baited breath to see if our offices in Singapore, Tokyo Geneva, Frankfurt, Paris, Madrid, New York and many more could see it. A few telephone calls established they could and a collective sigh of relief went up.

Since that day video has become a regular channel as part of Reuters internal communications strategy. Largely it’s been used to bring one-off interviews with senior managers to the notice of staff – to enable them to explain often-complicated strategy in a way that’s both personal and easy to digest. On occasions it’s also been used to make short tapes – usually not more that five or six minutes – about aspects of the company’s work. These have had good feedback from staff and have undoubtedly been popular.

Access to the videos has been on demand via our main corporate website, Daily Briefing, and have appeared on the site, usually back up by a text story and other information. As yet we haven’t made any attempt to launch a ‘channel’ with any sort of promise of continuous or near continuous programming. So far the emphasis has been on quality rather than quantity, but it seems likely we will turn over more material in time. This is made even more possible thanks to a software upgrade made by Cisco to the system last year. It’s proved a sensational bonus.

Since last spring we have been able to stream video live to our 13,000 desktops without any further modifications to the equipment. As a result, in July 2002 – a year after the original service was offered – we were able to beam our CEO’s results presentation to analysts live to our staff globally. Once again the system worked well and encouraged us to plan a more ambitious event last December.

Taking a meeting for staff in our London headquarters by our CEO as a base, we installed three cameras and encoded it on the spot – once again to go out live to our staff round the world. As an extra refinement, and to make it truly interactive, we invited our staff to send questions to our CEO during the meeting via e-mail and messaging, which were displayed on a screen in front on him during the meeting to answered on the spot. The exercise proved very popular with the staff that took part and we’re now looking at the possibility of similar events in other parts of the world, tapping into the Cisco system to send the images from wherever we generate them out to the rest of the Reuters world.

Separately we’ve also now done work linking standard video-conferencing cameras in our London headquarters into the global distribution system. This means individuals or groups can transmit live, theoretically to the entire company network and then archive their transmission for retrieval on demand later. If the broadcast is only designed for a narrow user group, details and a unique URL giving access to the broadcast can be mailed direct to the members of that group only.

In short, now that our system of Cisco servers have been installed globally it’s becoming obvious with each passing day that there are more and more uses that the network can be put to. At the same time a powerful return-on-investment case is being made. Virtual meetings and broadcasts to targeted groups can have a dramatic affect on cutting travel costs – always an issue in an international company.

Summing up, Reuters experience with its internal video system has been a good one to date. We’re still learning, there’s a lot more we want to do but I think everyone who’s worked on this project agrees it is exciting technology, it does work, and the potential applications are numerous. And, by the way, with good and imaginative technical backup of the kind we’re fortunate to have in Reuters, it doesn’t have to cost a fortune.

If you’re sitting in a medium to large company now and you’re thinking to starting something like this, my advice would be simple:

Think big (global) – it’s the only way to get the job done;
Gather good technical support around you;
Get buy-in from the top;
Set clear specifications and targets and make sure you stick to them (and everyone else does too).

But don’t take everyone’s advice. As I said at the start, in TV everyone thinks they’re an expert – but only a few really are.

Steve Clarke is head of internal media at Reuters. He can be contacted at steve.clarke@reuters.com