Industry leaders gathered in Dubai last month for a closed-door meeting to discuss some of the challenges around a MAM implementation and how best vendors can support them. Vijaya Cherian brings you the highlights Media Asset Management is a subject that has continuously challenged vendors owing to the complexity and uniqueness of the requirements at each broadcast facility. […]
Industry leaders gathered in Dubai last month for a closed-door meeting to discuss some of the challenges around a MAM implementation and how best vendors can support them. Vijaya Cherian brings you the highlights
Media Asset Management is a subject that has continuously challenged vendors owing to the complexity and uniqueness of the requirements at each broadcast facility.
Although discussed endlessly, it appears that most MAM solutions in the market so far have failed to adequately address the evolving demands of broadcasters, who not only have tonnes of undigitised tape without metadata but who also now work across multiple platforms and need a solution that is flexible and can be easily integrated with their existing workflows. Some vendors, of course, argue that they have addressed broadcasters needs but it requires agility and responsiveness as needs change.
This served as the premise for a high-level, closed-door discussion on Media Asset Management between some of the tech experts in the regional market and Avid, who sponsored the roundtable, hosted by BroadcastPro Middle East.
Joining us for the discussion were Saleh Lootah, Deputy General Manager TV & Radio Support, Head of TV and Radio Engineering at DMI; Dominic Baillie, CTO of Sky News Arabia; Robert Taylor, Programme Manager for Al Jazeera Workplace Transformation Project; Mark Billinge, CTO of OSN; Omar Alzoubi, Engineering Systems Senior Manager, DMI; Peter Van Dam, Senior Manager, OB Planning at Abu Dhabi Media Company; Laurent Tescari, Director of Product Development, E-vision, Etisalat; Bassem Maher, IT and Technical Manager, Takhayal Entertainment FZ LLC; and Craig Dwyer, Senior Director Global Centre of Excellence at Avid Technology.
The roundtable was moderated by Avids Director of Strategic Solutions, Peter Ennis, formerly Executive Director of Technology and Broadcast with Al Jazeera Media Network. Ennis, having worked in the regional landscape and being a respected name in the industry, encapsulated some of the pain points surrounding MAM in the region and engaged the tech heads to speak about their biggest challenges, with possible solutions to address them.
He began the conversation by sharing some of Ovums preliminary research findings, which Avid has also supported. The research, conducted among senior executives including CEOs, CTOs and tech heads within the broadcast industry, showed that 40% of the respondents highlighted poor integration of broadcast systems as the primary pain point in a MAM implementation.
The research indicated poor integration, constant search for different middleware formats, and the break in the chain if one element is replaced as significant causes of concern.
Likewise, trying to reduce the unit costs of media assets on site was cited as a significant point. This included storage and retrieval costs and optimising ROI in that asset essentially, being able to find the asset so it can be reused when necessary. With consumers now dictating how they watch content, broadcasters clearly have considerable challenges ahead of them.
THE TECHNOLOGY – Why implement MAM? Possible goals, challenges and solutions
As Al Jazeera has already invested in a MAM solution, Taylor pointed out that the implementation was based on the promise that our assets could be managed, accessible, controllable and available to the wider population at the network.
The history of Al Jazeera is that it is still partially tape-based. It has vast libraries of unique material to which the network had no access because we didnt know where it was and were not sure what each tape had. We were told MAM would be able to do that.
The second reason for investing in a MAM solution was because it could potentially manage everything automatically for everyone. The ingest people would ideally press a button and it would appear in the control room playout. That is the vision our news executives and senior executives had.
Peter Ennis, having overseen the project for a while at the network, added that the need to access and use any of that content across different repositories in different countries was a key requirement of the Doha-based network.
Mark Billinge of OSN, however, commented that his team foresaw deploying a MAM system at the pay TV networks facility as being hugely challenging.
Our perspective is very different from that of a news organisation, he clarified.
We still handle a lot of content through a fast-growing dynamic workflow and I guess thats one of the reasons we havent integrated an overarching MAM system. I think to try and overlay on top of what we have would be incredibly difficult and measuring that ROI and what it gives you is very hard to do.
He added that OSN currently finds existing MAM systems hugely restrictive given that the pay TVs current workflow, which has grown significantly to accommodate its evolving VOD and online platform requirements, already places huge demands on his team.
We have grown rapidly over the last five years and the workflows have changed enormously to be able to support the on-demand platforms whether it is Go, Play or the on-demand platforms on the set-top-box. I am not sure we would have achieved that in the ways we have if we had the limitation of a MAM system.
He added, however, that the network has attempted to create a bespoke workflow automation system that incorporates some of the elements of MAM.
We did it to automate some of the more complex parts particularly around on-demand like transcoding content and delivering content out to the various platforms. Having those automation processes took a lot of the headache out of the workflow. I am not saying that what we have is perfect but its better than investing hundreds of thousands of dollars in a classic MAM system which, at the moment, would be more of a risk than an advantage to the business.
This led the team to once again revisit what a MAM system should potentially do.
Peter Ennis took the lead here to explore what MAM systems have historically been like and how they have evolved.
Historically, it has been a system that was used to search for assets that you are storing and retrieving but it has evolved over the years to incorporate many of the other elements that customers want. But the primary question still is, how efficiently is your organisation using those assets and how easy is it to retrieve them?
He added that he has seen very inappropriate implementations of MAM with people wanting only some elements of it for their organisation.
This led to the question of whether vendors should really impose all four elements of a MAM system onto an organisation, or give it the flexibility to choose.
By the four elements, he refers to finding assets (storing and retrieving); integration of different systems; orchestration, which is moving the assets and delivering it in the correct format where it is needed and optimisation, which refers to using advanced reporting tools for business process optimisation to make best use of scarce resources.
Baillie of Sky News Arabia commented that the most important feature that every end user looks for in a MAM system is the ability to search and retrieve assets efficiently. However, he was sceptical about any traditional system being capable of doing this to complete satisfaction.
We looked at existing MAM systems a couple of years ago and saw that Googles search appliance was a lot more useful at finding content than traditional solutions. I worked with a broadcaster a while ago that has a huge archive of content that hasnt been managed properly. I know there are answers to that but it costs a fortune to digitise or even to log and get proper metadata from your existing archive. And I havent seen a MAM vendor that has looked into that. There used to be a vendor called Digital Smiths, which had a reasonably intelligent MAM solution.
Their MAM looked at what data I can get out of the content, and it also analysed the content. It took text from closed captions and looked at object tracking and had some good ideas. I havent seen anything like that again. I want to be able to install something in my station that finds the content, tells me what it is and helps me find something when I dont know what I am looking for.
Baillie mentioned that at the moment, Sky News Arabia uses the IP Director from EVS to undertake some of its functions.
We get the script from the agencies and it is ingested automatically. A librarian adds key words to it but all of this costs time and money and we shouldnt have to do it. We would be having a different conversation today if the audio script was transcribed automatically and served as rich metadata.
He added that Microsoft has already started doing something along these lines with Azure and the cloud.
You may start to see competition from the less traditional broadcast vendors on this front, he cautioned.
I would assume that with web-based TV gaining ground over traditional TV, Microsoft and Google and the way they manage their content are beginning to look extremely attractive, because then, the majority of content that is being consumed would be on their platforms.
In essence, end users have addressed the concerns around traditional platforms by either developing partial internal automated workflows or using their existing systems to perform additional functions. We gathered from the discussion that participants desired to better utilise the assets they have around search with richer metadata. They also wanted to be able to create that metadata on the fly rather than having people create it.
Dubai Media Inc. which, at present, is evaluating MAM options, pointed out that choosing the right MAM solution is a challenging process, as there are many MAM systems for different types of workflow.
Its not just about the technology; there is also a business process and workflow involved, Alzoubi stated.
I also think customers are still not able to define their requirements clearly and likewise, vendors are not able to cater to customers fully.
Moving from tape to tapeless and from linear to non-linear is a big transition that requires a lot of skills upgrade, and training can’t always achieve it. We need vendors to provide a MAM that can add value to each process in the workflow and also utilise fully existing resources efficiently and effectively.
CMS vs MAM
This led to a discussion on CMS systems and one participant queried if CMS and MAM systems were not duplicating functions.
OSNs Billinge commented that most CMS systems can indeed do 90% of what a MAM solution can do although he added that the pay TV networks existing CMS system left much to be desired.
I would like to see the CMS system go much deeper into the broadcast workflow and interact with the hardware and other systems in the workflow. Thats the direction in which we want to go, commented Billinge.
Craig Dwyer, who heads Avids Centre of Excellence, gently pointed out that while there were several similarities between CMS and MAM solutions, there were also significant differences in the way both technologies were designed.
We find that technologies are all fusing together now. But it is not as much about the technology as it is about the knowledge of the workflows, and implementing it and integrating it. If you talk to a CMS vendor, they will know more about digital distribution, web content management, servers, caching etc. That is their cult world. They understand the workflows, the matrix, the SLAs and so on. If you talk to us, it is a lot more about where those technologies meet mission critical news workflows, sports workflows and all the partners in integration that surround that and the training and workflow etc.
He added that any organisation with a group of skilled business analysts and developers, who understand the nature of the companys business and the technologies they require, can easily assemble solutions and adapt them to suit their requirements.
At the same time, he agreed that MAM solutions still have a long way to go.
We are using more open technology now, which allows for more elastic search across multiple glossaries, he said, alluding to the search function in a MAM solution.
Everything you want is not intelligently done now and not popping out like magic. We need to get there. If we were using more cloud technologies, we could start to do more intelligent background searching, algorithmic searching, pattern matching, big data analysis and so on. These things are coming.
Actually, I think Facebook and Google are putting a lot of their research into open source. You will see that a lot of what we are doing is to get MAM to a fully viable platform technology that can be deployed for libraries repeatedly and ensure integration. There is just a huge amount of complexity when you look at different vendors and the different integration points.
He said the big challenge was when people tried to layer this technology.
Everyone ends up with a custom configuration and end users are building all these anchor points that dont give you the flexibility you thought you were going to get.
It was agreed across the board that a proper search feature that is elastic as well as intuitive is the need of the hour.
Taylor seconded that adding that most broadcasters today require a Google experience, where the system is intelligent and reacting and can think for you and assist you.
He also recommended a different design approach.
“At every such deployment, you tend to see the engineers come in, plug in the boxes and then try to figure out how it will work, and then we have all this pain about the interface doing this and not doing that. These are core engineering mistakes. If you want to get investment out of your MAM, you need to learn to do it properly to start with.
Baillie also pointed out that vendors have unnecessarily complicated a system that could be easily engineered.
If you build a system from scratch, and move the archive into the centre of the system and start bolting the bits around that content and metadata, it would be a lot simpler, he recommended.
THE PROCESS – Who gains the most from MAM?
Dwyer added at this point that the clients, gaining the biggest advantages from MAM solutions, were those that had the capabilities themselves.
They may start with wanting a vendor to implement a system for some specific use cases and we may have some out-of-the-box solutions to take care of that. However, at some point, the organisation has to own the capabilities of the platform and when there is a core team that can work with the business, we see the greatest success. They dont want a vendor with a professional services team coming in all the time and doing their work. They want to be self-directing, reconfiguring the system, adding new workflows and so on.
Billinge stated that this was one of the main concerns at OSN.
Even if you get this system to work with this workflow, we change so rapidly a lot of the time. Its about each of the pieces in the chain accommodating change. If you do have a piece that is slowing us down, it becomes a headache for us. It depends on how flexible and dynamic the system is.
Baillie seconded this and added that in many cases, vendors lock end users down with their solutions.
We buy a system from you; we have to specify that system for you to deliver it properly. When you force me to lock it down, how can I potentially have a flexible system that can understand my content and make it available in a way that I want it available?
Peter Ennis agreed that this was the way vendors had to move to satisfy broadcasters needs.
The simplicity that you are looking for is you have multiple repositories, perhaps different codecs and you want to transcode them into a common house format, and then it is more universally shareable. We need to get to the stage where the system sorts through the content in a more intelligent fashion.
Challenges at different media houses technology and costs
Peter Van Dam of Abu Dhabi Media presented another problem. As a tech facilitating company that provides all the support not just for the state broadcaster but for external parties as well, they are constantly asked if they can change their business model.
We have content that comes in from different clients and each of them want to operate with a slightly altered workflow. That ability to reconfigure our system on the fly and deliver value to our customers is becoming increasingly important. We want to be able to define use cases and we want the system to deliver the capability to change use cases and adapt quickly to the changing environment in which one operates. In any case, we want to be able to deliver measurable deliverables against which we can justify a return on our investment. Then it becomes easy to justify the revenue.
This compelled the team to once again examine the sort of matrix each of them uses to determine the business viability of investing in MAM.
How do you look at the return on that versus buying an off-the-shelf system? Peter Ennis queried.
DMIs Saleh Lootah said an off-the- shelf system would never work.
It limits my usability of the brand and forces me to adapt to something I dont want to.
Rob Taylor questioned if there was indeed an off-the-shelf MAM.
Every MAM I have touched was taken from ground zero every single time, he pointed out.
Peter Ennis commented that Avid does have something pre-configured with standard workflows.
But when customers look at it, they say it is perfect for me but I need some tweaks and some changes. It can provide a starting point but not ground zero.
While some said it would be difficult to justify a MAM implementation in the light of the inability to monetise assets that can be neither searched nor retrieved, Taylor and Baillie raised the issue of licences.
Why does it have to be that expensive? Baillie queried.
Rob Taylor seconded that.
Why does a vendor feel that they have to charge more for the licences once the system has been bought? Why does one have to pay for every single person who accesses it if an end user has paid for the system, the core and everything else?
Speaking from the vendors side, Dwyer justified this model stating that the model for enterprise systems is based on concurrency.
If you have 300 people, you need 100 licences. We are running a business, and we are investing a significant amount of money in R&D. These products are aimed at a relatively small market with maybe a few thousand customers. The only way to sustain the systems and make them robust and have engineering solutions to support it is to charge a premium. We are, of course, trying to find the right balance with the foundation configurations. We are now able to get to the few hundred thousand dollar change. That is a significant change from several years ago.
Cloud a possible solution?
Laurent Tescari, who works with E-vision, commented at this point that one of the ways to address the cost issue was to look at cloud options.
I am unable to justify millions of capex. I prefer to work with cloud partners on flexible business models, and I realise I have to work with new players to ingest content and process it, but at least, it would make it a more cost effective model for a department like ours that has only around 50 people.
Tescari went on to add that ROI should always be linked to the business and not to the technology, which is essential to get moving. He commented that the technology was like the wheels of the car and questioned what the consequences of the car not moving were.
He further justified his choice of a cloud model stating that with IP and 4K looming on the horizon, he would have to again make significant changes to his workflow if he made an internal investment.
Having a cloud vendor who can handle that would be a smarter thing than having to take it on internally.
Bassem Maher of Takhayal said that having an efficient search function would have a positive impact on revenue within his organisation.
We have clients who want our content and we just cant find it. We spend half of the money we get from the vendor to just retrieve the content. What we need is the ability to search for our content correctly end to end.
Takhayals CMS web site is currently doubling up as a makeshift MAM solution. It stores all of the companys metadata.
We are indexing our content using our scheduling system without moving to central storage. We are justifying this by using our CMS and our scheduling system as our indexing system, he explained.
Peter Ennis agreed that there was only one thing worse than not having the rights to a piece of content.
It is not being able to access the content to which we have the rights, he said, quoting an old industry colleague.
He added that it was hugely crucial that end users have the capability to monitor content to which they had the rights and know when they were expiring to automate the process of sending an email to the scheduler closer to the time of the date of expiry, so that the rights can be renewed.
There may be limitations around rights management, but at least, you can ensure you are exploiting fully the potential of the content to which you have the rights.
Most of the end users agreed that building a MAM system from scratch would be a massive undertaking and not something they would attempt.
On this note, we took a short break for networking and returned to an analogy from Peter Ennis, who said that large technology deployments are like a stool with three legs the tech leg, the process leg and the people leg. While the first is often the easiest, the last is the toughest.
If you get any of these legs wrong, the stool will fall over, he cautioned.
THE PEOPLE Evoking change
Saleh Lootah agreed, adding that it is the people who decide the success of a tech implementation.
Citing the issues he has commonly witnessed, he said people are reluctant to move out of their comfort zone.
It is difficult to get people to adapt to change. With every new introduction, we face so much resistance. With MAM, the resistance would be huge.
Most of the end users commented that they had tried to devise strategies to incentivise staff to adopt the workflow.
In DMIs case, Lootah says conflicting workflows between two departments served as the catalyst for change.
When we gave the creative department a luxurious file-based environment, they loved it but it put them off when they realised they still had to take their output onto a tape and give to the library. This created conflict and eventually, the library realised that they needed to transition to a file-based environment.
Billinge seconded this by adding that people and efficiency are some of the key measures of ROI.
Peter Van Dam agreed that this was a significant challenge that he faced too.
Would the adapt or get out concept work in most government entities?
The looks on the faces of some of the end users instantly signalled that they would have to work with the same people.
At this point, Craig Dwyer suggested a strategy that was used at a Scandinavian broadcast facility.
The broadcaster had implemented MAM but couldnt get its journalists to put in metadata. At the same time, the broadcaster had also introduced a VOD platform for the stories. Now all the journalists did want to see their stories on screen and see which ones got the most hits. The library manager hit upon a great idea. He stipulated that the journalist can only publish in the VOD platform if they have their metadata.
Essentially, they just moved the incentive a bit. If it was just more work, the journalists didnt want to do it. But the moment they saw the benefits, they were really happy and were enthused to do it. Being able to think about how to incentivise the right behaviour or put it into your appraisal matrix is something to think about, he suggested.
Closer home, Rob Taylor said Al Jazeera experienced success by taking its staff on a journey from the first day instead of leaving training for the end.
Engage your users from Day Zero of a project. It is all about the communication to the end user both ensuring we are getting information coming in from them and that they are being taken on the journey from Day Zero. Not towards the middle and certainly, it wouldnt have worked towards the end.
Thats something I have learnt over a lot of years of doing this. You bring them in at the beginning and take them on the journey. It doesnt solve all the problems but it takes the number of people who are resistant down to a very small number. It takes effort and investment and requires a lot of skills upgrades. You have to take them away from their world of tape and introduce them progressively to a world that is not tape. So while they may not have used the tapeless system yet, they will at least be familiar with the terms and the concepts.
Peter Ennis, having worked on the programme himself, elaborated on the project.
In this case, we did a skills audit primarily around the technology bit and also around the operations. We looked at everybody we had within a very large technical organisation and matched the present level skills against the time when the technology would be implemented. We looked at what skills would we need at point x in the future to adequately support, implement etc.
I felt sick when I saw the results but instantly, you could see the size of the problem but then, it also gave us the opportunity to address it.
The end users agreed that in most instances, a technology is fully deployed and commissioning has been completed before users are trained on the system.
While some private players argued that they were not obliged to take staff on this journey, those working in government entities said that unlike corporate entities that were driven by profit and loss, they worked in a different environment that had to accommodate cultural sensitivities.
Taylor, however, was quick to point out that the method of taking people on a journey from Day Zero is advantageous.
Once you take them on that journey, they themselves become excited and want to know when it is all going to come together. Take them on the journey, he urged.
Baillie added that people dont need training to use Google or Facebook because they have made the user interface easy and intuitive.
Make it easy and its easy to take people on that journey, he said, commenting that an intuitive MAM would be a very desirable solution.
The discussion slowly tapered towards whether 4K and the availability of ultra HD content would drive the uptake of MAM in the business.
Baillie immediately pointed out: If you design your systems properly now with the archive at the centre, it would not matter what format you transitioned to. It impacts it now purely because of the way the systems are designed.
How can vendors add value?
Peter Ennis concluded the conversation by asking participants how vendors could help add value to end users.
Key requirements were the need for more open systems so that end users were not locked into a specific vendors environment; the need to simplify operations and costs; and the need to make MAM more intuitive, so it is intelligent enough to understand the inputs from different islands within a system and adapt.
Avid was also asked to look at metadata more closely.
Metadata will become more valuable than content in the next few years. Focus on metadata, Baillie urged.
Dwyer added that Avid was taking more components and making them horizontal, to ensure they were made available across more platforms.