Push-to-Talk

Push-to-Talk: the licence, the radio, and the reason the two are priced apart

Push-to-talk is the oldest interface in professional communications and the newest problem in industrial connectivity. The interface has not changed since the 1940s: press, speak, release. What has changed is everything underneath it. A pressel that once keyed a VHF repeater now keys a session on a 3GPP mission-critical platform, running over a private LTE core, authenticated against a corporate identity provider, and recorded for evidentiary use.

That migration is what makes push-to-talk worth analysing as a market rather than as a product line. The service is cheap; the coverage that makes the service usable is not. A push-to-talk licence is priced per user per month and behaves like software. A radio network is priced per square kilometre and behaves like civil engineering. Every question about this market — who wins, where margin sits, why a buyer chooses one architecture over another — resolves to which of those two cost structures the buyer is actually paying for.

What this segment covers

Pheonix Research treats push-to-talk as four layers that are bought together and priced apart:

• Service and platform — push-to-talk over cellular (PoC) licences, mission-critical push-to-talk (MCPTT) platforms, dispatch applications, recording and location services, whether sold as a carrier subscription or a self-hosted platform.
• Devices — rugged PTT handsets, intrinsically safe and ATEX-rated units, vehicle mounts, body-worn cameras and accessories.
• Radio infrastructure — private LTE and 5G radio and core, TETRA, P25 and DMR systems, distributed antenna systems and leaky feeder for underground coverage, and the backhaul that connects a remote site to anything else.
• Integration and managed service — network design, installation, systems integration, spectrum authorisation support, and the multi-year service contracts that increasingly wrap the other three.

Excluded from scope: consumer messaging applications, general enterprise unified communications, and public mobile voice tariffs that are not sold as a push-to-talk service.

Why the layers diverge

The economics of the four layers are not related. Device supply is a global, competitive, low-margin hardware business with Asian manufacturing concentration and short product cycles. Radio infrastructure is an oligopoly of a small number of original equipment manufacturers with long qualification cycles and site-specific engineering. Service is a subscription annuity whose price is anchored to a public mobile tariff the customer can see. Integration is fragmented, labour-intensive and local.

The consequence is that a buyer rarely faces one supplier. They face a device vendor, a radio OEM, a carrier or platform provider and a local integrator, each with a different margin structure and a different appetite for risk. The layer that carries the highest price is usually the one with the least competition, and it is rarely the layer the buyer spends the most time negotiating.

The coverage constraint

Push-to-talk is a coverage product before it is a feature product. Where public cellular coverage exists, a PTT service is a software decision and can be bought by the seat. Where it does not — a brine field at 3,700 metres, an underground copper mine, a 60-kilometre pipeline corridor — the buyer must first build the network, and the push-to-talk decision becomes a network decision with a capital budget attached.

This is why the segment’s centre of gravity has moved from the operator to the operator’s customer. Mining companies, ports, utilities and energy operators are now among the largest private deployers of cellular radio in the world, and the push-to-talk service that rides on those networks is frequently the smallest line in the business case that justified them.

How to read the reports in this segment

Pheonix reports in this segment separate published statistics from modelled allocations, and say which is which wherever a figure appears. Regulatory dates and thresholds are stated from the instrument itself; subscriber counts, per-user economics and segment splits are modelled from site-level assumptions unless a national statistical source publishes them. Each report carries a basis-of-estimate section naming the assumptions that carry the most weight, so a reader can test the conclusion rather than accept it.