A Texas personal injury attorney calls and says the phone records are already in hand. The opposition answered the request for production, the production arrived as a PDF, and the theory is that the defendant driver was on the phone when the collision occurred. Then the file comes over, and what is actually in it is a monthly statement showing calls rounded to the minute, no tower fields, no inbound message detail, and no session records at all.
That is a billing record. It is not a Call Detail Record (CDR). Understanding the difference between billing records and Call Detail Records is one of the most consequential pieces of CDR analysis in crash litigation, because the two documents sound like the same thing and look similar on the page, but behave very differently under analysis.
I put the question “What’s the Difference Between Billing Records and Call Detail Records?” to attorneys on LinkedIn and offered four differences: time precision, cell site fields, completeness of SMS and data events, and usefulness within the crash window. Every one of those answers is correct under the right facts. Which one controls depends on the carrier, the retention period, how the request was drafted, and what the case needs to prove. Below, I walk through each one, separating what the records directly show, what they only suggest, and what cannot be determined from that source alone.
A definition first. A billing record is the accounting artifact a carrier generates in order to charge a subscriber. A Call Detail Record is the network’s event-level log of activity associated with a subscriber line, generated for switching, routing, and internal operations, and produced in varying formats depending on the carrier and the request. One document exists to justify a bill. The other exists for network maintenance and improvement. Both are system-generated, which is part of why both are useful, but they are built for entirely different purposes.
Option 1: Time Precision, Minute Versus Second
When this is true. This difference matters in nearly every crash case. Billing statements commonly express call start times to the minute and durations in rounded up to the next minute for billing purposes. A CDR typically carries start time and duration at second-level precision, along with a distinct end time to the second, or a duration time to the second, from which the end can be calculated.
What the records directly show. A CDR entry showing a call beginning at 14:32:07 with a duration of 143 seconds directly establishes a connected voice event running to roughly 14:34:30. A billing entry showing “2:32 PM, 3 min” establishes only that a chargeable call began somewhere inside the 2:32 PM minute and was billed at three minutes. This leaves a possible end time of the call between 2:34:01 PM and 2:35:59 PM, a 1 minute and 58 second window.
Limits and caveats. Billed minutes are not elapsed minutes. Carriers round up and rounding conventions differ across carriers and plan types. A call lasting 65 seconds may bill as two minutes. A call lasting 121 seconds may bill as three. If the collision is time-stamped by a 911 CAD entry at 14:34:12, the billing record cannot tell you whether the call ended forty seconds earlier or was still connected at impact. The CDR frequently can. This is the practical version of the two-minute problem: the ambiguity at the tail end of a billed call is often wider than the entire window counsel is trying to reconstruct.
What cannot be determined from either source alone. Neither document establishes who was holding the phone, whether the audio ran through a hands-free system, or whether the driver was looking at a screen. Connection is not necessarily conduct.
Option 2: Cell Site and Tower Fields
When this is true. This difference controls when the case turns on where a vehicle was, not just what the phone was doing. Billing records never carry cell site location data. CDRs, depending on the carrier and the scope of the request, may include a cell site location data and sector information for the subscriber end of a voice connection. A sector is one side slice of a tower’s coverage, typically one of three facing different directions.
What the records may suggest. Cell site location data along with a sector field associated with a call indicates which cell site and sector served that connection at that moment. It can be consistent with a phone being somewhere within that cell site and sector’s coverage footprint. In a trucking case where the driver’s account places him thirty miles away on a different route, a sequence of sector records inconsistent with that account carries real impeachment value.
Limits and caveats. Sector data is not GPS. Coverage areas vary with terrain, antenna height, horizontal beamwidth, vertical beamwidth, and network load. Sectors overlap, and, on occasion, a phone can connect to a cell site that is not the geographically nearest site. Network conditions, load leveling, and handset behavior all influence which site serves a given connection. I will say a record is consistent with a handset being within a coverage area. I will not say a tower record places a tractor-trailer in a specific lane, at a specific speed, or at a specific intersection.
What cannot be determined. Cell site location data and sector information in CDRs generally reflect individual connections, not continuous tracking. Gaps between calls are not evidence of being stationary or moving. If the case genuinely requires location, that is a separate analysis drawing on device artifacts, telematics, or ELD data, and it should be scoped that way from the start.
Option 3: Completeness of SMS and Data Events
When this is true. This difference controls when the theory involves texting or app use rather than a voice call. Billing statements often show only message counts or bundled allowances, and they may omit inbound messages entirely. They commonly express data usage as aggregate usage for the billing cycle, with no event-level breakdown at all.
What the records directly show. A CDR produced with messaging detail can show individual SMS and (sometimes) MMS events with timestamps, direction, and the other party’s number. Data session records, where a carrier retains and produces them, can show session start and stop times and the number of bytes up and down. These are the call logs, text logs, and data logs that make a pattern of communication analysis possible.
Limits and caveats. Content is the recurring disappointment. Carriers generally do not retain the text of SMS messages beyond very short (a few days) operational windows, and MMS content retention is mostly nonexistent. Messages sent through iMessage, WhatsApp, Signal, or Facebook Messenger do not traverse the carrier’s messaging platform as SMS at all. Those travel as data through the internet, which means they appear, if at all, as undifferentiated data activity. Recovering that content requires the device itself or a lawful cloud production, not a carrier subpoena.
Data sessions deserve a separate warning. A handset with background email sync, automatic app updates, and location services enabled generates data activity while sitting face down in a cupholder. A data session concurrent with impact does not establish that the driver was viewing the screen. Treating it as proof of active use is a common overreach, and it is the kind of overreach an opposing expert witness is happy point out.
What cannot be determined. Which application generated a session, whether the screen was illuminated, and whether the subscriber or a passenger was the person interacting with the device. Attribution is its own evidentiary problem, and it is a real one in commercial vehicle crash litigation, where drivers may carry both a personal phone and a company phone, and where a company handset may be managed through an Mobile Device Management (MDM) platform that keeps logs of its own and can make a phone extraction difficult.
Option 4: Evidentiary Usefulness in Crash-Time Analysis
When this is true. This is the answer that matters when the question stops being academic. Can the document in your file support an admissible, defensible opinion about a window that may be only sixty to ninety seconds wide?
Consider a Texas commercial vehicle case. A tractor-trailer rear-ends a passenger vehicle on a highway outside Waco. The CAD entry places the first 911 call at 4:41:38 PM. Defense produces the carrier’s billing statement, which shows one outbound call at 4:38 PM lasting four minutes. Read plainly, that call could have ended anywhere from roughly 4:41 PM to 4:42 PM, which straddles the impact. Plaintiff’s counsel argues the driver was on the phone. Defense argues the call was over well before impact.
Neither position is supportable on that document. The billing record cannot resolve it. The corresponding CDR, if obtained, might show a start of 4:38:12 PM with a duration of 188 seconds, placing the end at approximately 4:41:10 PM, 28 seconds before the 911 call. That is a materially different case, and it can move in either direction depending on what the actual numbers turn out to be. Layered against ELD and telematics data showing the last recorded speed change, and against an event data recorder download, the CDR becomes corroborating evidence inside a defensible timeline rather than a standalone assertion.
Limits and caveats. Time zone handling is not a footnote. Carriers commonly produce records in UTC, sometimes in a fixed zone that does not observe daylight saving, and occasionally in local time without saying so anywhere in the production. A one-hour normalization error destroys a crash-window opinion. CAD systems, EDR modules, dash cameras, and ELD platforms each keep their own clocks, and those clocks drift. Every timeline I produce documents the source zone and the deviation applied for each data stream, because that is the first thing a competent opposing expert will test.
What cannot be determined. Even a flawless CDR establishes network events, not human behavior. It shows that a call was connected. It does not show that the driver was distracted, and it does not establish causation. That inference is built from a pattern of communication across multiple sources, and it is argued, not assumed.
Tips for Attorneys
- Involve a cell phone records expert early, before the subpoena goes out. The cheapest hour in the entire engagement is the one spent reviewing the request language before it is served. Once the retention window closes, no amount of motion practice recreates data the carrier no longer holds.
- Request Call Detail Records by name, with the fields enumerated. Ask for event-level records including start time, end time or duration expressed in seconds, direction, originating and terminating numbers, device identifiers, and cell site and sector identifiers where retained. A request for “phone records” invites a billing statement. Each carrier has its own language. Contact me at ben@braveinvestigations.com for cell phone record preservation letter examples with the specific language for the carrier in your case.
- Preserve early and separately. Carrier retention for detailed records is materially shorter than for billing data, and it varies by carrier and record type. Send the preservation letter before you know whether you will need the analysis. Follow up with new preservation letters every 90 days until you can send a subpoena.
- Corroborate rather than rely. In commercial vehicle cases, pair CDRs with device extractions, ELD, telematics, EDR, and dash camera data. More data is almost always better.
- Request subscriber records alongside the CDR. Account name, activation date, device identifiers, and authorized users all bear on attribution, which matters when a driver carries both a personal and a company phone.
- Retain the native format. Ask for the records in .txt or spreadsheet format rather than PDF. Experts work in Excel, and converting a PDF into a spreadsheet can be difficult, especially with AT&T records.
- Have the return reviewed before the deposition, not after. Field definitions, carrier-specific codes, and missing record types are far easier to address while the window to supplement or move to compel remains open.
Conclusion
All four answers are correct, and choosing among them
follows a straightforward sequence. Ask what the case needs the records to show.
If it needs impact time activity or location information, billing record will
not get you there. If it needs to test a driver’s account of where the vehicle
was, cell site location data shows coverage areas rather than position. If the
theory is texting or app use, content of messaging, and data events, you should
expect content to reside on the device rather than with the carrier. And if the
practical question is whether the document in your file can support an opinion
at all, that is the fourth answer.
Billing records are accounting documents that happen to
mention communications. Call Detail Records are network event logs that happen
to be useful in court. Both can be authenticated. Only one is generally built
for the precision that crash litigation demands.
How Brave Investigations Can Help
If you have a carrier return in hand and are not certain
what you actually received, we review call detail records, text logs, and data
logs for pattern of communication analysis, interpret subscriber records and
subpoena returns, analyze cell site records with the, and build timelines and
cross-party communication matrices for personal injury litigation and
commercial vehicle crash litigation. We also review mobile device extractions,
including outsourced examination reports, and support examination protocol
drafting and defensible reporting.
Ben Bierce, Brave Investigations, Texas Department of Public
Safety license number A28412101. Reach us at ben@braveinvestigations.com
or through the contact form on this site.
Related Reading