BAFTA View is a VOD service, exclusively used by BAFTA members, to review Film, Television and Games Content entered into the BAFTA Awards. Previous to 2020 the distributors sent DVD screeners directly to members, which had a net cost of up to £50k per film. BAFTA has been hosting and streaming Television programmes for over 10 years, however, the security requirements required by the studios for pre-release content had been a sticking point for BAFTA.  This issue was solved in 2019 when Screener-hub, an API integration platform, was built by BAFTA Tech. Screener-hub is middleware which enabled studios to host films on the platform of their choice, ensuring their security requirements were covered and retaining ownership of the asset. The film is then streamed to pre-authenticated BAFTA members within BAFTA's software ecosystem. Screener-hub was an instant as it addressed the concerns of the studios and enabled access for members and with limited cinema screenings in 2020 due to COVID-19 the service provided a solid base for members to screen films. With Screener-hub the biggest piece of the puzzle missing was a sophisticated VOD front end where all the video content and entry metadata could be listed together to replace the then curent viewing workflow.

BAFTA Members have been voting online since 2002 through a secure voting platform designed and managed by BAFTA. Viewing was fully introduced into the Voting Site in 2013 for Television, Games, Childrens and regional awards, but by 2020 the design had become dated and the requirements had grown in terms of displaying increased volumes of metadata collected in BAFTA Nucleus, the awards management system built and manged by BAFTA Tech.

During the development of Screener-hub the decision was made to build a new front end for Awards viewing which would become BAFTA View. UI Centric, an external design agency, were commissioned to design a front end VOD which would look modern and represent BAFTA and flexible enough to support complex workflows required by the BAFTA Awards and by judging the result you can see how well they succeeded. My role, as Product Manager, was to turn the designs into an MVP for the Film Viewing which was due to open in September 2021. For this challenge I had 9 months and a part time development team of 3, as the developers also worked across other projects.

At this stage I do not want to underestimate the complexities of delivering such a project. I inherited some beautiful and feature heavy Figma designs, which were great, but they did not always consider the technical challenge of importing the metadata from Nucleus, BAFTA's Awards entry system and database, to displaying on the BAFTA View front end. Many of our users keep suggesting we use an external VOD such as Netflix, Disney + or one of the many competitors, however, they do not consider that these systems do not face a key challenge that BAFTA does, and that is the collection, review, storage and then display of huge amounts of metadata. Sure, if Netflix commission a programme one of the key components will be for the production to complete documentation and to supply assets in a specified format. That is very easy to achieve if Netflix will not pay the production company until all deliveries have been completed. Production companies have processes to deliver this. BAFTA in a position to pay production companies X millions, and the standard of date for each broadcaster/distributor etc. is different. When BAFTA receives data it comes from a multitude of sources, in a variety of formats and standardisation is not an option. BAFTA therefore has to populate the VOD with data and images that are different and are not keyed into a central database in a standard. Even collating a synopsis is a huge task and the range of this one item of data is different. Netflix can control its voice, BAFTA needs to display what the entrant has supplied so long as the synopsis meets the guidelines.

Taking the above explanation into consideration my first challenge was to see how the designs translated into reality. Do they match awards processes? Can Nucleus display certain data in the required format as stipulated in the designs? For stakeholders and members it is very easy to request a feature. The designers can consider the UI, crate beautiful design and position the feature in the design but its an entirely different question to deliver this in code. For each feature I needed to consider if we have the data stored in a method that can be applied. For example, a feature which is helpful to many users to to be able to sort films by runtime. This sort option had been suggested and is included in the designs. However, the Nucleus database does not store the film runtime. As stated before BAFTA does not receive many of the video assets so BAFTA's systems do not have access to video file metadata and as a result determining the length of a file is impossible. There is a question on the entry form which asks about film length - this is more to determine eligibility and the conclusion was to use the data in this field as provided by the entrant. The problem is that the answers are not accurate. Many entries are created and submitted before the final version of the film has been edited. Eligibility requires the run time of the film to be more than 70 minutes, if a user enters 90 minutes and the end runtime is 100 minutes then this was not previously a problem. But using this data for a sort filter does present a problem as users are expecting a 90 minute film and when they watch it the film is 100 minutes. BAFTA Admin are unable to confirm this as how can they know the full length of a film while it is being edited and although they could recheck there is limited resource and time to action these processes.

The point is to illustrate the complexity and thought that is behind every decision and feature within the site. Usually the answer is a compromise, we could have reviewed the sort function but I felt that it would be more useful to display some inaccurate runtimes amongst the mostly accurate in the long term than remove the feature entirely.

To further detail the complexity of the integrated systems I've included a system integration diagram to give you an idea of the relationships and complexity. And the relationship of all these systems had effects on the final implementation of BAFTA View.

After the initial design review I realised that some key workflows had been missed for the Television and Games Awards. Working with the designers I requested some additional pages and features that were then incorporated into the designs.

While the designs were being reworked I had two main focus points. The first was to capture all of the components that had been created within the designs and to create Jira tickets for the developers to build the component library. I'm sure any Product Manager that has translated designs into a product understands that there is a level of interpretation from both the Product Manager and Developers. My role here was to ensure that the components matched the designs and that the developers implemented the designs as specified, not how they interpreted. This initially resulted in a slow start to the build but once the developers understood my vision they quickly changed the results.

The second focus was to create user stories for the required features. Working with the product owner and stakeholders I reviewed which features were required for MVP, which features were not practical to deliver at this stage and features to reserve for future releases. I found that during the scoping process many stakeholders had added suggestions, many of which were complex, but there was little evidence to suggest users would utilise the features. Due to time limitations my approach was to deliver a clean site with a basic level of features. I wanted to conduct in-depth user research after the site had been used for an awards season. Allocating resource to building features that there is no proven use case is expensive and can be a waste of time. My previous experience has taught me that once users access the site their requests will change. And through the post MVP research this approach has been proven correct.

Before starting on building such a complex project many decisions had to be made. With a limited development resource and a fixed timeline I needed to deliver a working VOD system with API integration to three separate systems. Therefore I looked for elements where I could reduce resource.

One key area was how the front end of BAFTA View would be administrated. Awards admins already have multiple logins to multiple products and I was hesitant to create yet another a new site that admins users would need to login to. Not building an admin site would also reduce the workload on the developers and help us to achieve the delivery time. I therefore decided not to build an administration access and BAFTA View would be managed solely using functionality in the in the existing BAFTA Voting Site and Awards Management System. I decided to build an api integration mapping screen which would connect the metadata fields with the relevant location in BAFTA View. The result is that menus and page structure would be hard coded but Admins can specify the data set that is placed within a given mapping.  All elements that would need to change, such as categories, awards and entry information was already configurable in either the Voting Site or Nucleus.

The user authentication process was another area that was changed form the designs to save development time. BAFTA had plans to introduce an SSO for members in-order to navigate through different products, such as membership area in Salesforce communities, BAFTA Events (a ticketing solution I also manage), BAFTA Voting and now BAFTA View. This has now been launched. The key requirement of BAFTA View is that members should only have viewing access to content that they can vote for and this is allocated at category level per user. The voting site manages these credentials, importing the user rights from Salesforce records. BAFTA View therefore would already use this data to ensure users had the correct viewing privileges. It was a small step to decide that the Voting site could be used to authenticate the users with the existing 2FA as opposed to building the solution from scratch. This also ensured that members completed mandatory surveys and courses before logging reaching BAFTA View.

The third big decision was to outsource the entire help section to WordPress. The designers had created beautifully detailed designs for a detailed help section which the Membership Team could then manage by creating bespoke instructions. The decision to not build administration access had been made and creating a bespoke CMS just for one section of the site seemed an unnecessary undertaking. After some research WordPress was a logical solution for the problem. Why build something new when it already exists? WordPress has the added benefit of being able to purchase a theme and then applying the BAFTA View styling. I found a theme that presented itself as Help/FAQ site which contained similar components. After one of the developers installed WordPress on an AWS server I set up the theme and menus and another developer supplied the styling. The teams supplying content for the members have a cost effective fully editable solution where they can create new sections and articles as they require which needs little maintenance.

The core of BAFTA View is the API. The API pulls all of the required data from BAFTA's systems into BAFTA View where it can then be used by the front end to populate the pages. BAFTA Voting and Nucleus are built around the concept of Awards and Categories and due to members viewing constrains so is BAFTA View. This meant that we could compartmentalise the data and supply the required metadata for each Award. The challenge was that each different award type required different data sets.

After analysing the pages provided by the designers I created a table of metadata required for each page, per award. I then broke this down further to create a list of mappings. For example, a film synopsis was required on two separate pages then a synopsis mapping was created in the api. This mapping would then be used by the developers building the front end pages to display the synopsis data on the relevant pages.

Instead of building a bespoke mappings page in Nucleus it was decided that the question building functionality could be used to server this purpose. I broke down each mapping type so they it could only be used with a specific question type. The synopsis mapping can only be applied to a text area input field and a video a video upload input field etc. Once set this was applied to each relevant question type to enable admins to specify the data.

This was further complicated is that each award has its own specific parameters. The Games Awards, for example does not have a main viewing video, but TV and Film Awards do. I reviewed all of the requirements across the different award types and we were able to attribute all the different requirements to 5 controls:

• has video
• has category page
• primary judging media
• has runtime sorting
• has download button

Within these controls I could specify all of the unique parameters require, so if an Award has category page set to yes this triggers a number of different options and functionality that is no available if this is set to no and vice versa. These options are listed in the voting site against each award meaning that admins have full control over how to set up BAFTA View for each different award. This also resulted in not having to build a new mapping control page which again reduced costs and time.

The challenge was to build a fully responsive, fully accessible mobile, TV and desktop friendly website application which would adapt depending on the number of awards, categories and entries available to view at any one time. Again, to iterate this is not a usual requirement of a standard VOD. Netflix for example is curated and populated by an algorithm. A typical VOD service never has to deal with an empty page and the same metadata is supplied for each item, potentially with the exception of film vs series. The next step in the process was to build the pages to match the designs once the components had been completed taking into account for the scenarios specified.

I used Confluence and Jira to list all of the different components that were required to build each page. The product is feature heavy and as a result the tickets contained detailed information describing the build. Many of the pages required significant tweaking as again the designs displayed a page that was full of content but in reality the pages needed to work with minimal content as well as full metadata sets. This is something to consider when commissioning designs. Make sure you ask your designers to provide examples of pages with minimal content as there will always be use cases where pages need to work with limited content. I know that I will do so in future.

The final step was to build the functionality. Having beautiful pages and an API is great, but not practically useful if there is no functionality on the front end. This process was a long process of trial and error, taking painstaking time to evaluate the designs alongside testing on BAFTA View. Functionality on one page would often change functionality on another page an ensuring the whole workflow was seamless was a difficult task. User stories were created in Jira and tickets created and recreated. Testing would highlight new functionality that was not immediately evident from the designs, especially in terms of user navigation but slowly the site became fully operational.

Testing was another intensive process. Setting up the test environment took time. Nucleus is very good at exporting the entry forms but not the entry data. I had to hand create many example entries so that we could test how the metadata date pulled through and was displayed on the website. This helped to populate the site and provide a better understanding of the final look and feel of the site which required more tweaking.

I am immensely proud of BAFTA View and the huge amount of technical work carried out by the small team on a limited budget and short timeframe. Feedback, which has been positive, have highlighted areas of improvement which will be implemented ahead of September 2022. Interestingly many of the features that were specified by stakeholders and then removed for MVP have not been mentioned - this is why sometimes its important to deliver less than more for an initial iteration.

I've included some images of BAFTA View, as for security reasons it is not possible to link to the site. I'm particularly proud of this project as it delivered all of the aims in the first year and the user response, especially to the designs and interface has been very positive. The next step is to implement the design on the Voting site so create a coherent user journey for members and the current difference in design between the sites is a little jarring.

Above are some examples of the different list screens. There were initially four examples in the design but I removed one as it was too similar to the poser view. The three screens are designed to support different users browsing habits. Some users are very visual and others prefer lists of entries.

Below are some examples of the entry reviewing screen. This screen is hugely complex as it needs to list all the categories where an entry is listed, video and image content, supporting statements, download files, and general metadata. There is a feature for users to claim game codes where a video is not used. All of the content changes per entry and category entered, so if an entry is not successful in a given category then fields need to be removed. This makes the whole process extremely complex to setup.