How to Shortlist and Compare Persistent Video Platforms for Post-Production

Why build a shortlist by elimination?

Screening and comparing are different jobs. Screening asks whether a candidate is eligible at all and gives a yes or no. Comparing asks which eligible candidate is better and gives a judgment. Run them together and a strong demo can override an unmet requirement, which is the most common way facilities shortlist a vendor they can’t actually use.

Screening first also protects the scarcest resource in the process: time from post supervisors, assistant editors and technical staff during periods when they’re already busy. This guide covers the buying process. For what to look for in the platforms themselves, including fidelity, re-encoding, color and content security, see our guide to persistent video collaboration priorities for post-production.

Which screens come first?

Security obligations your productions actually carry. On studio work, contracts often reference the MPA Content Security Best Practices, which the Trusted Partner Network maintains along with its assessment program.¹ Read what your contracts require. Some require vendors to hold a specific assessment, which is a hard screen. Many set rules for how material is handled, which several kinds of platform can meet, so ask each vendor how. Applying the strictest standard you’ve heard of, instead of the one you’re bound by, removes suitable candidates for no reason.

Support hours. Post work runs in the evenings and at weekends. Ask for coverage hours in writing, and who is reachable when a session fails at 9 p.m. with an executive watching.

Onboarding time. Decide the longest your productions can tolerate between requesting access for an outside vendor and that vendor joining a session, and remove candidates who can’t meet it.

Exit terms. How long material is kept after a show wraps or the contract ends, and what deletion is committed to, because productions often require material to be destroyed on a set date. Vague answers here become expensive exactly when a production wraps. Ask what can be exported too. It matters most for anything that will exist only in the platform; if your masters, notes and decisions live elsewhere, a platform without export can still pass this screen.

Three finalists is right for most facilities. Two gives no real comparison; four or more usually means screening was incomplete. If honest screens leave only two, go with two instead of relaxing a screen to reach three.

Where do the candidates come from?

Start by knowing what kind of supplier you need. Vendors from different origins use the same words, and “persistent room” means something different to a company that began in file-based review than to one that began in live video transport.

Then build the long list from people, not search:

• Peer facilities running comparable work. Ask whom they use, whom they stopped using, and what went wrong after signing.
• Your own outside vendors. Editorial houses, visual effects vendors and colorists join other facilities’ sessions constantly and know which platforms are pleasant to join.
• Industry associations. The Hollywood Professional Association is a non-profit member association connecting businesses that provide creative and technical expertise, tools and infrastructure for professional content creation, distribution and archive.² Its members and events map who serves this market.

What should the information request ask?

Ask for evidence, not descriptions. “Do you support persistent rooms?” gets a yes from everyone. A written session specification, security documentation, a support coverage statement and exit terms give you something to compare.

Ask what “persistent” means in their product: does the room keep its media, layout, participant list and permissions between sessions, or only its address? And ask for any latency figure on a stated basis. The IETF defines streaming latency as the glass-to-glass time from an event to its playback on the viewer’s device, including encoding, decoding and buffering, not just network transit.³ A figure with no basis isn’t a specification.

Keep it short. A long questionnaire sent to eight vendors creates work for everyone and rarely changes the outcome. A focused one gets answered properly by serious vendors instead of being handed to a proposal team.

How do you make demos comparable?

Assume a vendor’s own demo is optimized: their content, their infrastructure, small audiences. So standardize everything before the first session: same material, same participants, same connection, same length, same things recorded. And drive it yourself, with your footage on your connection and your people joining from their usual locations.

Run each session in three parts: a short vendor orientation, a run led by the facility, and a deliberate failure segment in which a path drops, someone joins on a weak connection and attendance exceeds the plan. The third part is where candidates separate. Run at least one session in the evening at realistic scale, when your upload is already busy with renders and transfers, and see whether vendor support picks up.

Record the same measurements every time:

• Time to set up the room
• Time for a guest to join from outside
• Observed delay on a directed action, such as “stop there”
• Visible artifacts on your hardest material
• What happens when a weak-connection participant joins mid-session

Agree the weighting of each dimension in writing before any demo. Weights set afterwards tend to justify a preference that has already formed. And include the assistant editors and post coordinators who will run sessions every night; their view of the fiftieth session matters more than anyone’s impression of the first.

How should cost be compared?

Vendors price different things: named users, viewers, concurrent rooms or streams, storage, quality tiers. Headline rates aren’t comparable. Take a real show at a real stage, count the rooms, sessions, participants, hours and media kept, and ask each vendor to price that month. Then price your busiest foreseeable month, and ask what happens when you exceed a tier. Rankings often change between the average month and the peak.

Add the costs outside the subscription, such as your own upload capacity, encoding hardware or software, and the time to standardize room setup. They vary with each platform’s architecture.

How should references be checked?

Vendor-supplied references will be satisfied customers, so ask them specific questions: what went wrong in the first ninety days, and how did the vendor respond? A reference who can’t recall any friction either had an unusually smooth start or isn’t being candid.

Find at least one reference the vendor didn’t supply, through peers or your own outside vendors. And find one long-running user, because a facility two months in can’t tell you how a platform behaves at eighteen.

What happens between shortlist and decision?

Run one real session on a live production with the leading candidate. A real dailies or editorial review with real stakes shows more than another demo, and most vendors will support it if asked.

Run the security review in parallel, not at the end. Waiting for a favorite to emerge before starting adds weeks and sometimes reveals the favorite was never viable.

Get commercial and exit terms in writing from every finalist, not just the leader. Comparable written terms are your leverage, and asking only the front-runner gives it away. Protect yourself the same way whatever the vendor’s size: retention and deletion terms, export rights where material lives only in the platform, notice of material change, and a shorter first term if doubt remains.

Then set a decision date and keep it. Shortlists go stale within weeks as staffing, productions and pricing change.

How to run it

1. Write the four screens from your actual contracts and working hours.
2. Decide which kind of supplier you need, and ask each vendor what its product was built to do.
3. Build the long list from people: peers, your outside vendors, industry associations.
4. Send one short information request asking for evidence, a definition of “persistent” and a stated latency basis.
5. Weight the dimensions in writing, then run identical demos with a failure segment and an evening session.
6. Price a real month and a real peak with every finalist.
7. Check one independent and one long-running reference, run one real session, get written terms from all, and decide on the date you set.

Coresee makes a persistent, browser-based video collaboration workspace, one of the kinds of supplier this guide describes.

Published November 9, 2026