The radio layer, drawn so a planning meeting can actually use it.
Pipeline Flowgrid is a small analytics practice. We take the counters, drive-test traces, and incident logs a Georgian operator already holds, and we return maps, annotated worst-cell lists, and a written note the room can argue with.
We do not sell a login, a licence, or a wall of tiles. The work is a study, a sitting, and a pack you keep.
A study ends with paper you can take into a capacity meeting.
Most requests come from RAN planning, a NOC lead who needs an incident reconstructed, or a regulatory pack that has to show more than a single nationwide KPI. The flagship engagement is a radio-access performance study for one market or one technology layer.
A typical pack includes cluster maps at busy hour, a ranked list of cells with the counters that put them there, a short written finding on each theme (congestion, coverage holes, backhaul constraint, seasonal load), and one briefing of up to half a day in Batumi or at your operations floor.
We do not change live configuration. We do not sit in your OSS as a standing vendor. If you need that kind of retainer, we will say so plainly and point you elsewhere.
Other studies the same desk takes on
These sit beside the radio-access study when a market needs a narrower cut — backhaul, coverage, an incident, or a GNCC-facing KPI pack.
Backhaul and core traffic review
A reading of microwave hops, fibre tails, and the packet-core interfaces that starve a healthy radio sector during the hour that actually matters.
Coverage and interference mapping
Drive-test traces and predicted coverage, drawn against the live site list, so a hole on the ridge is not treated as a capacity problem on the boulevard.
Incident reconstruction
A written timeline of a named evening: the cells that failed, the cells that held the load, and the counters that contradict the first war-room story.
Tourist-season load, mountain shadow, and a coastline that fills on Fridays.
Batumi’s radio layer does not behave like a flat capital. Summer weekends pile onto the boulevard and the port approaches; inland Adjara still has ridges that swallow a sector. A nationwide average hides both. We spend our time on the cluster, the hour, and the site list you actually operate.
Operators we sit with already export Ericsson, Huawei, or Nokia counters. We do not ask you to rip those out. We ask for a clean site list, a known-outage calendar, and a sentence about the meeting the pack has to serve — a capacity freeze, a spectrum refarm, a GNCC filing, or a board note after a bad night.
Client notes, not scores
They spent the first week arguing with our site list, which was irritating and also the reason the congestion map finally matched the cells we actually transmit from. The briefing in Batumi was two hours, not a slide theatre. I still wish they had flagged the microwave hops a day earlier — we had already booked the backhaul vendor by then.
The incident pack reconstructed the Thursday evening drop without asking the NOC to sit through another war-room recap. Names of counters, times, and the two cells that held the load.
From the study desk
A busy-hour counter is not a congestion verdict
Why a single PRB figure at 11:00 on a Tuesday misleads an Adjara capacity meeting — and which hours we ask for before drawing a map.
Look at the site list before you talk about a refarm
Renamed cabinets, leftover 2G identities on microwave hops, and why a spectrum conversation collapses if CGI and the hop inventory disagree.
What the Batumi boulevard does to a summer cluster
Tourist-season load is not a footnote in Adjara. Here is how we put Friday evening next to a quiet Wednesday without turning the pack into a postcard.
Request a study
Write with the market, the technology layer, and the meeting the pack has to serve. We reply from the Gorgasali Street desk on working days.