An answering machine (or “Voice Mail”) is an automatic system that answers incoming calls and plays recorded messages. This system can inform about the status of the called number (for example, “the subscriber is currently speaking” or “the dialed number is turned off”), provide instructions (IVR) or play another automatic message, after which it goes silent, giving the interlocutor a chance to provide feedback.
In our work, we divide answering machines into two main types:
- Cellular network answering machines are technical messages indicating that the subscriber is unavailable or roaming. A feature is that the message is played without simulating picking up the phone (sending a “connect” phone command from the subscriber to us to establish a two-way voice connection), after which it forwards the call to voicemail, which the connected call center agent then hears;
- Voice Mail or PBX Answering Machine – its feature is that it first sends “connect,” and then plays the message. In the case of a PBX of some organization, this can be an entire script with instructions on how to select certain options, contact the right department or employee, and so on.
In call centers, when agents call clients, they often encounter answering machines, which can impede productive work. One way to solve this problem is to use answering machine detection technology. The system recognizes when an answering machine is on the line and appropriately drops the call, allowing agents to focus on speaking with live customers.
How the system recognizes answering machines
Answering machine detection or AMD technology is a technology that allows call centers to reduce the number of calls directed to voicemail or unavailable numbers.
How does it work? If your customer call list includes non-existent, disconnected phone numbers where a standard answering machine or voicemail is activated — the system recognizes such a number during the dialer and won’t switch the call to the agent.
Often, an answering machine is identified by a specific sequence of words, such as “The called subscriber cannot take your call right now. Please leave a message after the tone.” A real person would say something like “Hello,” and the system would transfer this call to the agent.
The problem is that answering machine detection is not an exact science and this causes difficulties:
- Time delay while the system checks if the subscriber is an answering machine, which can cause confusion on the part of the client and lead to the call being terminated. That is, the client answers the call and hears silence;
- False positives, where the system mistakenly identifies a person as an answering machine, which can lead to the loss of a potential customer as the call will be dropped.
Nevertheless, this technology is still relevant and often discussed. There are debates about its usefulness and the problems it may cause. We will consider the aspects and benefits of answering machine detection in our article.
Disadvantages of answering machine detection
Answering machine detection is an important topic for call centers, because false positives (when the system identifies a person as an answering machine) are the most common problem. Nonetheless, there are many services and solutions of varying effectiveness. For example, some SIP communication providers offer to detect answering machines during dialer customers.
To achieve 100% accuracy in detecting answering machines, it’s necessary to drop only such calls, but this is practically impossible, as detection methods may sometimes miss live subscribers or fail to recognize voicemail.
The Marketing Association (DMA) claims that no more than 3% of calls within 24 hours can be dropped. “Dropped call” is a call where a connection to a live person is established, and then a disconnection occurs – drop.
Moreover, calls not transferred to the agent within the first two seconds are also subject to this limit. That is, the client should hear the agent’s voice almost immediately after picking up the phone, rather than listening to silence while the system verifies whether it’s a real person or not.
So, if your detection system cannot reliably perform within two seconds, it’s better not to use it and consider alternative strategies.
Although answering machine detection can be useful for saving agents’ time, it’s not always effective. Dropping calls can affect future relations and customer loyalty. Slow detection and subsequent delay can irritate the client and even worsen their impression of your company.
Dropped calls and answering machine detection technology (AMD)
Recently, answering machine detection (AMD) technology has attracted widespread attention, especially in the context of automated calls. The British regulatory agency Ofcom expressed serious concerns regarding the use of AMD, noting its connection to the increasing number of silent calls. Their statement contains recommendations aimed at limiting the percentage of dropped calls to a level of 3%.
The crux of the problem is that, despite developers’ assurances, AMD cannot always detect answering machines, especially considering modern digital technologies. Using AMD can significantly increase the number of dropped or silent calls due to a large number of cases where this technology erroneously classifies a person as an answering machine.
So if your service doesn’t operate in real-time or your prediction algorithm isn’t set up as it should be – you won’t be able to adhere to this 3% limit of dropped calls.
Smart dialing: strategies for reducing the number of missed and silent calls
It’s quite simple, why annoy a potential client or customer before you’ve spoken to them? Consumers have been suffering from intrusive, silent or, worse, endless dialing attempts for more than a decade. The results were harmful to the entire industry, many subscribers blocked their numbers from unwanted calls, and number-checking websites became in demand. The number of people you can contact by phone decreased, meaning that each number is considered precious.
Now is the time for a more refined approach to dialing customers, rather than sheer persistence. Call centers should revisit priorities and focus not just on the volume of calls made, but on improving the efficacy of their activities. This includes smart use of contact history, as well as skillfully creating call lists.
Tags or hashtags can be a convenient tool for identifying numbers on which an answering machine is operational. Instead of making repeated calls to these numbers, you can immediately mark them accordingly in the system. This will prevent unnecessary attempts to call non-existent numbers and clients with voicemail.
Moreover, tags and hashtags can be used to plan subsequent contacts with clients. For example, you can set a period after which the next call will be made to a number with an answering machine. This will better organize the dialer and take into account the temporal characteristics of clients.Finally, make sure that, upon reaching a live person on the other end of the phone line, agents will be fully prepared to engage with the client. Aim to personalize interaction with clients, listen attentively, and focus on the quality of the services provided.
What to replace answering machine detection services with
There are alternative strategies that allow for more systematic and targeted dialing of clients, increasing the response rate and reducing dropped calls.
- Analytics and segmentation: Collecting and analyzing data about clients helps better understand their needs and preferences. This allows for creating more accurate and targeted call lists. Segmenting the customer base enables addressing different groups of clients with different messages, increasing the likelihood of successful contact;
- Contact history: Maintain a history of contacts with clients. Records of previous communications and requests allow agents to be informed during subsequent calls. This creates a more personal and trustful communication with clients;
- Integration with CRM systems: For example, in CRM oki-toki an agent can indicate that this number has an answering machine or system message “Number does not exist” and there’s no need to call it anymore;
- Optimizing dialing time: Use data and analytics to determine the optimal time for dialing different customer categories. This can significantly increase the chances of successful contact.
All these strategies are aimed at increasing the efficiency of dialing, reducing dropped calls, and improving interaction with clients. Choosing a specific strategy depends on the goals and characteristics of your business, but in today’s environment, where clients are increasingly demanding in terms of service, using alternative dialing methods can be more productive and offer long-term benefits.
Practical advice for call centers
Working with answering machine detectors (AMD) can be a challenging task, and improper use of this technology can lead to the loss of potential customers. Here are some tips for your call center for more effective call management and reducing the number of dropped calls:
- Do not use answering machine detectors if it’s not clear that it will bring significant benefits. For example, there’s no need to use them for B2B campaigns or when dialing a database of numbers where the percentage of such calls is low;
- Don’t forget about the delay between answering the call and connecting it to the agent (usually at least 1-2 seconds). This happens because all calls must go through the AMD algorithm to detect answering machines with some degree of accuracy. The delay effect irritates many people who might hang up after the first couple of greetings. Therefore, you may notice an increase in the number of dropped calls by the very people you’re trying to talk to;
- Assess false positives for each campaign with AMD enabled. The key point is understanding that AMD, while useful, is not always the best solution. It’s important to find a balance between minimizing dropped calls and ensuring a more effective and pleasant interaction with clients;
- Use alternative strategies to optimize the dialing process. Don’t rely solely on the answering machine detector; pay attention to more effective methods, such as optimizing call time and creating personalized contact lists. A systematic and targeted approach to dialing clients can increase the response rate and reduce the number of dropped calls.
Answering machine detection technology has its pros and cons. Time delays and false positives can create inconvenience for both clients and agents, but AMD technology continues to evolve and improve. In the future, with the development of artificial intelligence and machine learning, the answering machine detection system will become more refined, reducing the number of errors and increasing the accuracy of answering machine detection, whether to use it now is up to you.
