Robocall Radar #3: Short & Silent Calls

Published on: April 30, 2024

Welcome to the third is our “Robocall Radar” series where we are examining what can be learned from our RRAPTOR robocall surveillance system. In the first couple of instances we profiled the nature of many of the calls that RRAPTOR was receiving, including analysis of call contents, technical errors observed, and relentless calling patterns.

Here we’re looking specifically at the hordes of short and silent calls that RRAPTOR receives.

YouMail, one of the most authoritative sources of robocall statistics, has been telling us for years that there are about four BILLION robocalls smothering Americans each month. YouMail categorizes the calls as alerts, reminders (including payments), telemarketing and scams – based on what the message from the caller.

But if the caller doesn’t say enough – or anything at all – YouMail isn’t able to categorize the call nor count it as a robocall.

Most of us get plenty of robocalls, and we also routinely answer calls where nobody is there. We get missed calls on our phone and no message is left.

Of the millions of calls RRAPTOR has received, MORE THAN HALF (up to 60%) don’t contain enough audio to categorize. This is consistent, week after week. Most of these calls come from the same set of service providers that facilitate the robocalls we do recognize.

Our tentative conclusion: The actual number of robocalls – calls from automated calling platforms, regardless of what message (if any) they deliver – could be TWICE what YouMail tells us – approaching 10 BILLION in peak months.

A RRAPTOR Refresher

A reminder of what RRAPTOR,  our system for answering and analyzing phone calls, is and does.

RRAPTOR’s pool of thousands of telephone numbers (most listed on the National Do Not Call Registry) are answered by the RRAPTOR answerbot, which attempts to engage the caller as might a human call recipient. The call is both recorded and transcribed. Keyword analysis of the transcription is used to characterize each call.

Further, RRAPTOR captures and analyzes the STIR/SHAKEN information for the call. Not every call is signed, but most are. Thus, we gather a detailed picture of who originated the call, as well as its path to termination at our system. All the information we collect is held in a database where it’s available for various kinds of analysis.

Taken in the context of our entire pool of numbers, RRAPTOR is able to provide a reasonable statistical analysis of robocall activity impacting the public as a whole.

We believe the calls received by RRAPTOR provide a representative sample of robocall activity impacting the public as a whole.

A Match Game

Keyword/phrase analysis of the call transcription allows us to build an understanding of call patterns that broadly reflects robocall campaigns. When a call matches a known phrase or keyword we assign it to a category. For example, calls purporting to be from the Publishers Clearing House Sweepstakes or the Social Security Administration are definitely suspicious. Calls referencing these phrases are flagged as “Suspect.” All that really means is that we’ve been able to in some way characterize the nature of the call as subject to the TCPA or other regulation, or at least likely unwanted, and potentially worthy of investigation.

RRAPTOR receives many calls that are not in any way suspicious. We see a significant number of calls where a called party is returning a missed call from a phone number that is unknown to them. This often happens when robocallers are using a calling number other than their own (“spoofing”), and happen to use a number that is currently in the RRAPTOR pool.

Many other calls are flagged as “non-suspect.” For example, calls with alerts about school closings, weather alerts, or prescriptions ready for pickup. It’s clear that these are calls that were most likely desired by the intended recipient. There is likely no nefarious intent, even if the caller is not being diligent about maintaining their calling list.

Remember, virtually all the phone numbers in the RRAPTOR pool are on the National Do Not Call List. Conceptually, RRAPTOR should be receiving no robocalls at all. Yet, it receives many thousands of calls each day.

The Short, Silent Type

Almost half the calls answered by RRAPTOR are very short in duration and/or have no sound at all. We can only make an educated guess as to why this occurs. It could be someone honestly misdialed a telephone number. Rather than saying “Sorry, wrong number” they simply hang up. It may be the result of a predictive dialer that makes an outbound call, but does not have an agent available to take the call once it’s answered.

These calls can’t be characterized as suspect on the basis of the contents of the calls – because there is little or no content. That said, we still have the metadata about the calls to help us seek out patterns, and hopefully gain some insight. In many cases, thanks to STIR/SHAKEN, we usually know the identity of a voice service provider that facilitated a particular call.

The following table lists the top 10 companies signing short & silent calls to RRAPTOR between April 1-29.

The table above shows RRAPTOR received 1,417 calls signed by Sipphony over the period from April 1-29. 100% had no transmitted audio. 671 were completely silent, and in 746 they hung-up prematurely. None of the calls were matched to “suspect” categories since there was insufficient content to be evaluated.

The calls were to 511 of our numbers, from 700 different ANIs. Almost all the RRAPTOR numbers called were on the DNC list.

Calling 1,417 times in 29 days, that’s one call every 30 minutes, 24 hours a day. All silent.

To telephone someone, just out-of-the-blue, is an imposition. It’s a demand for immediate attention that tears them from whatever productive activity they might have been doing at the time. It expresses disregard for the called party’s time and attention. It’s hard to imagine that this is intentional on the part of the caller, so something is apparently malfunctioning at their end.

Somewhat similarly, 91% of the 1,564 calls signed by Technology Innovation Lab LLC were short, and 85% of 4530 calls signed by Jaintel International were short.

Unfortunately, it costs an originating carrier (or their client) so little to do this that they simply don’t seem to care.

At Scale

Zooming back out to the broad view, RRAPTOR received over 10,488 short and silent calls from just these ten sources. RRAPTOR has a relatively small pool of phone numbers compared to the entire universe of telephone numbers in use in North America.

This suggests many millions of short and silent calls were actually placed. That’s a tidal wave of interruption.

Silent and short calls are a huge part of the robocall scourge. Callers and providers need to urgently address this issue. Poorly designed systems and implementation bugs need to be corrected. And if callers are intentionally hanging up (because, perhaps, they think they have detected voicemail), they need to alter their practices to leave an explanatory message rather than continue to frustrate call recipients by making the reason for the call mysterious.