Skip to content

Fighting Robocalls: What ARCEP Decision 0881 Means for Telecom Professionals

· Callr

Fighting Robocalls: What ARCEP Decision 0881 Means for Telecom Professionals

Phone numbers are a shared resource, and like any shared resource they are easy to over-use and hard to govern. France’s telecom regulator, ARCEP, took on exactly that problem with Decision 0881: a sweeping update to the national numbering plan that also tightened the rules around automated calls and text messages. If you run a contact center, send voice or SMS campaigns, or operate a network, the decision changes the ground you stand on. Here is a clear-headed look at what it does and why it matters.

Managing a common good

In telecommunications, providers constantly share standardized resources: phone numbers, URLs, IP addresses. When everyone optimizes for their own short-term gain, the resource degrades for the whole group, including the people exploiting it. Economists call this the Tragedy of the Commons, and the numbering plan is a textbook example.

The framework that governed French numbering had aged badly. ARCEP set out two goals with Decision 0881: foster innovation and new uses, and improve how the scarcity of numbering resources is managed.

ARCEP is the French regulator for electronic communications and postal services. Like the FCC in the United States, its job is to regulate the market and keep competition fair among providers and other players.

The decision tackles looming number shortages in some regions and adds restrictions on geographic numbers. But it goes well beyond a numbering update: it also introduces new protections for end users, and corresponding obligations for businesses, against automated calls and messages.

A few facts set the scene:

  • The previous framework dated back to 2001 and 2005, and no longer fit how networks were used.
  • Demand for numbers kept rising, driven in part by machine-to-machine (M2M) traffic.
  • It was technically possible, and too easy, for businesses to conceal the caller’s identity on automated calls.

1. The two sides of a call

Regulating telecommunications is a balancing act. The regulator has to reconcile what businesses need with what citizens and customers expect, and the two pull in different directions. Businesses are also better organized and better funded, which gives them more weight with regulators.

The tension is real, and businesses sometimes have to be protected from themselves. Without any guardrails, the market drifts toward another tragedy of the commons: companies flood the network with automated calls and texts until people do whatever it takes to escape the noise, from call-blocking apps to swapping numbers to abandoning voice and SMS altogether.

Rather than framing businesses and citizens as opponents, it is more useful to find the overlap. In telecom, their interests are aligned up to a point, and that shared middle ground is where good regulation lives.

Don’t forget the third party: operators

There is a third side to every call: the operators. They have commercial relationships with both end users and the businesses that use their network, legitimately or otherwise. Since operators earn on both sides, it is tempting to assume they have little incentive to fight robocalls. In reality, operators have to manage shared infrastructure like a scale: excess on one side eventually drives excess on the other.

1-A. Innovation that serves businesses

Often, what businesses want and what consumers want line up. Automation is a good example.

  • For consumers, well-built automated support means faster answers, available any time of day or week, and sometimes the relief of not having to talk to a person at all. Done right, automation does not replace human agents; it frees them to handle the edge cases where they are genuinely needed.
  • For businesses, automation brings consistency. Every interaction follows a defined path, which matters enormously for brands that care about their image, and because it happens on a digital channel it becomes traceable and measurable end to end.

The detail is where things go wrong. Some companies treat automation purely as a way to cut headcount and minimize support costs. That approach is common, and it usually damages both brand perception and customer satisfaction. And so far we have only discussed inbound support. The full picture also includes outbound campaigns, which is where most consumer frustration concentrates.

1-B. Rage against the machine?

For consumers, automation feels very different depending on direction. Inbound automation, where the user reaches out to a service, is usually welcome: it improves accessibility and typically sits alongside human support.

”Human!”

Even in that best case, frustration shows up. People learn tricks to escape rigid systems: shouting “agent” or “human” at an IVR, pressing 0, hammering the pound key, anything to reach a person better able to help. Entire resources now exist to help callers find a company’s number and bypass the menu to a live agent.

Outbound campaigns (robocalls)

Outbound automated calls and texts drive the bulk of end-user frustration. The techniques have multiplied: recorded-voice calls, SMS blasts, ringless voicemail. As the cost of a minute or a message kept falling, the conditions for a flood of automated activity were perfect.

A handful of tactics concentrate the anger:

  • Flash calling: ringing a number once so the recipient calls back.
  • Obfuscation: making a call look local when it actually originates abroad (more on this below).
  • Repetition: once a number enters a database, it tends to be shared widely. Some people receive five to ten automated contacts a day, and even carriers themselves are guilty of the practice.

1-C. Do we still need phone numbers?

One more piece of context. Communication has changed fast. Over-the-top (OTT) messaging apps, which carry conversations directly over the internet and bypass operators, have grown enormously in both users and usage.

OTT messaging apps let people communicate directly over the internet, bypassing telecom providers.

Some of these apps are still tied to a phone number; others identify users by email or username. On top of one-to-one messaging, community platforms like Discord and Slack have become primary channels for large groups.

Living without a phone number used to be impractical. It is now entirely workable, and for younger generations it is the norm: they keep a mobile plan, but most of their usage is data-based, and they rarely need to share a number at all. There are now plenty of ways to chat, call, or video-call anyone in your network.

With that context in place, we can look at the decision itself.

2. Decision 0881 at a glance

Decision 0881 is a large package; the original document runs well over 100 pages. If you operate a contact center or run automated voice or text campaigns in France, it is worth reading in full. Rather than covering every item, we will focus on the decisions that capture the spirit of the text.

2-A. A tighter framework around caller and sender identity

Caller Line Identification (CLI), or Sender ID, is the number or string shown on the recipient’s phone. Businesses used several tricks to mislead recipients by spoofing it, most commonly displaying a local number to lift pick-up rates even when calling from abroad.

The decision (Article 7.4) gives a precise definition of a “modified” caller ID:

The caller line identifier is considered modified when the number displayed on the end user’s phone cannot be used to call the caller back.

It then sets conditions to regulate the practice. The two most significant and novel ones:

  1. Two-way requirement: if you use a number to call or text a prospect, that prospect must be able to reach you back on the same number for as long as the campaign runs.
  2. Territoriality: when using geographic or non-geographic numbers (for example French mobiles 06/07 or geographic 01–05 numbers), the caller must be able to guarantee the agent is genuinely calling from France (Art. 7.4.2.b).

2-B. A new way to regulate automated calls and texts

ARCEP (Article 7.5) lists the main nuisances tied to automated campaigns and sets a framework to govern them.

Restrictions on automated communications

The section is dense, so here are the points that matter most:

  1. Automated dialers are explicitly named in ARCEP’s list of nuisances, though their precise legal status remained unclear even after we asked the regulator directly.
  2. New rules limit which number types may be used for automated campaigns and impose transparency on the caller. Automated systems can no longer use “territorialized” numbers (with the exceptions below).
  3. Three core limitations are defined, and they reveal what kind of automated traffic ARCEP considers beneficial on the network.

“Territorialized” numbers include:

  • Geographic numbers: 0[1–5] XX XX XX XX in France.
  • Mobile numbers: 06/07 XX XX XX XX.
  • Non-geographic numbers (distinct from mobiles): 09 XX XX XX XX, not tied to a region.

As a result, automated traffic is by default confined to one number type: long virtual mobile numbers, used mainly for automated and machine communications, including M2M.

The three exceptions:

  1. Small scale: systems calling or texting five distinct numbers or fewer.
  2. Balanced texting: automated SMS systems that receive as many texts as they send, or more.
  3. Balanced calling: automated calling systems that receive significantly more calls than they place.

The traffic ARCEP wants to encourage

Here ARCEP introduces a useful principle: automated communication should be balanced between inbound and outbound. The logic holds up. Why throttle a high-volume service that users voluntarily reach out to? A very low inbound-to-outbound ratio, on the other hand, is a reliable signal of spam.

Notably, the balance test is more lenient for text than for voice, which reflects ARCEP’s intent to support helpful automated messaging while cracking down on robocalls.

The decision is explicit that calls placed manually by agents, without an automated dialing system, are not treated as emitted by automated systems (Decision 0881, footnote 33).

The principle has clear limits, though. Two cases illustrate the problem:

  1. Critical alerting: a seismic monitoring system that warns local populations of a life-threatening event generates intense outbound traffic over a short window. It breaks the balance principle while literally saving lives.
  2. Debt collection: this work is mostly outbound by nature, since reaching a debtor often takes several attempts. The principle could brand a legitimate service as illegitimate.

2-C. Enabling new uses

ARCEP also made several moves to support innovation:

  • Make it easier to use multiple numbers on one device, so a single subscription could carry both a personal and a professional number routed to the same line.
  • Allow SMS and MMS on geographic numbers, which had been restricted to non-geographic numbers in France.
  • Make it easier to use long virtual numbers for IoT devices, including for users outside France.

Decision 0881 was adopted in July 2018, with the changes to automated calling slated to take effect in August 2019 to give providers time to adapt.

With the substance clear, here is how it plays out for the industry.

3. An insider’s view: a conversation with Callr’s CEO

Reading a new regulation is one thing; understanding its real impact is another, and that is where industry experts earn their keep. We sat down with Taoufik Zagdoud, CEO of Callr, the French smart voice services provider operating since 2009, to talk through Decision 0881 and its consequences.

What was your first reaction to Decision 0881?

Taoufik Zagdoud: As a VoIP carrier, our position is interesting. Legacy carriers have historically resisted extra regulation because they see it as a business risk. Our view is more balanced. We recognize that a strong legal framework is both necessary and good for the future of the industry.

We also invest heavily in our customer relationships so we can share what we know and help them adapt. The best practices are well documented; what we still need is to convince businesses that preserving mobile as a high-quality, responsive medium is in their own interest. We regularly publish guidance to push adoption of those practices.

There is no reliable way to filter traffic in advance without treating everyone as a suspect, so we work the other way around. We monitor our network closely and make it as easy as possible to report abuse, through a dedicated channel: abuse@callr.com.

Will it be enough to contain spam and abuse on the networks?

TZ: My experience tells me it won’t be enough on its own, but it is still a necessary step. The text defines and regulates concepts and technologies that were living in a legal grey area, and it is forward-looking, paving the way for new uses such as dedicated ranges for machine-to-machine traffic.

That said, neither the form nor the substance fully solves the problem, and here is why:

  1. Wishful thinking. Setting a principle to ban a practice is easy. Enforcing it is not. Regulators tend to assume technology can bridge principle and implementation, and here that is unrealistic. What even counts as “automated traffic,” and how do you detect it automatically? With telecom virtualization, you essentially cannot. There are so many ways to spoof identity or disguise automated traffic as human-initiated that any technical control will be highly game-able, and gamed.
  2. Automation is not the enemy. The decision reads as if automation were inherently bad. Plenty of automated use cases provide real, sometimes time-critical value: emergency alerting, non-profits helping homeless people find food, informational services.
  3. Real-world consequences. Most call centers serving developed markets are set up in lower-wage neighboring regions. They serve major banks, insurers, and telecom providers, so changes to telecom rules have a tangible effect on jobs in those countries. Even in a magical world where you could block all automated calls, demand for human agents would rise, pushing operators to chase ever-lower costs elsewhere. That race to the bottom is not a good outcome.

How is the industry reacting to the crackdown on automated calls and texts?

TZ: Among peers I hear real concern about the future of automated campaigns and contact center work. A crackdown on predictive dialing, if enforced, would have a major impact, and some go as far as saying it could jeopardize the whole sector.

I don’t share that view. It reminds me of the pre-GDPR panic: enormous expectations and worry about what would happen, followed by a far more moderate reality. With GDPR, many businesses found exemptions, and even those who couldn’t were not especially worried, because enforcement in the first year was very limited relative to the number of reported violations. Since GDPR alone did not curb phone harassment, ARCEP added the restriction that mobile CLIs (06 or 07) can no longer be used for automated campaigns. The pattern repeats elsewhere; the US has stiff robocall penalties on paper, but collection of those fines has lagged far behind.

The bigger pressure may come from consumer advocacy. The fear of a public backlash led by consumer groups can do more to clean up practices than the letter of the law. France’s La Quadrature du Net launched legal action against several large tech companies the day GDPR took effect, backed by thousands of citizens, and it produced results, including a substantial CNIL fine. GDPR’s biggest merit, in my view, was reminding consumers they have rights over their data and can assert them.

La Quadrature du Net is a French NGO that defends users’ privacy online.

TZ: We have offices in France and the US, so we follow both ARCEP and the FCC closely, and since we serve customers worldwide we track major changes from many countries. When we built our voice and text campaign tooling, we drew on FCC and FTC rules, and we generally aim to exceed the legal minimum. For example, we handle STOP messages at the platform level, so the feature is available in every country we serve, even where it is not legally required.

SMS traffic tends to be more tightly regulated and better enforced than voice. Several countries impose strict rules on text, including the United Arab Emirates, where shortcode use and campaign content must be approved in advance, and China, where routes are cut when spam is detected. As a rule of thumb, where authorities respond swiftly and forcefully against the operator, practices improve quickly, because operators themselves filter the unwanted traffic.

The voice-versus-text gap comes down to the nature of the traffic. Text is easier to analyze automatically, which simplifies spam detection, and the widespread STOP mechanism makes abusive behavior easy to spot when a company keeps contacting people who opted out. That is why flagging systems are easier to build and maintain for text.

France isn’t alone. What approaches do other countries use, and what works?

TZ: Over a decade in the field, I have seen two approaches deliver meaningful results.

The first, used in the United States, combines stronger laws, heavier enforcement, and larger fines. Tightening all three at once made a real dent in abusive and misleading calling.

The second is simpler and more direct: automated calls and texts proliferate because they are cheap, so some countries raised voice and SMS prices to make them less attractive. Rates can be raised further on inbound international calls to limit foreign call centers’ ability to dial in. This worked well in parts of Europe, notably Switzerland, Belgium, and Italy.

More radical models exist too. In the United Arab Emirates, marketing campaigns must be declared in advance to the regulator, and the operator is held directly responsible for any unlawful traffic on its network. That forces operators to be careful about what they carry, and it has measurably reduced abuse. Emirati operators now offer extensive resources to help business customers comply.

References and further reading

For a deeper look at the decision and the broader robocall problem:

  • ARCEP draft decision for public consultation on relaxing geographic restrictions for 01–05 numbers and increasing user protection (April 2019).
  • “The Source of Robocalls,” Om Malik’s blog (April 2019).

In French:

  • Décision n° 2018-0881 du 24 juillet 2018 de l’ARCEP établissant le plan national de numérotation et ses règles de gestion (Légifrance, July 2018).
  • “La réglementation de la numérotation change profondément,” Frédéric Foster, Lexing (September 2018).
Book a demoGet started for free