Is ARPU still a reliable metric?

In the telecom industry many AxPU metrics have been proposed and/or used to assess the business performance of a given telecom market and/or operator. However ARPU (average revenue per user) remains the most strategically regarded and widely accepted metric. It is commonly calculated by dividing aggregated revenues by the average number of subscribers who theoretically generated those revenues during a predefined time period – typically a month as adopted by most telecom carriers. More interestingly, focused ARPU figures are calculated for specific services in a bid to identify the largest sources of revenue generation, as they are calculated according to diverse factors such as geographic location, user age, occupation, and income.

Jamoussi believes it is time to contest assumptions such as low-ARPU markets are unprofitable, while high-ARPU ones are

A related metric to ARPU is ARPPU (average revenue per paying user), which is calculated by dividing aggregated revenues by the average number of subscribers who paid some amount of money, yielding a figure that is significantly greater than ARPU. Also related to ARPU – but less used – is AMPU (average margin per user), which is calculated on the basis of net profit rather than total income, and which is the difference between the cost of serving a user and the revenue that user generates. Thus the greater the AMPU the greater the profit. In this regard, it is noted that some telcos have started shifting their focus from ARPU to AMPU in order to maximise their returns as niche markets become saturated.

Today, as telecom markets reach saturation and competition aggressively erodes revenue streams, the challenge of securing targeted revenues with a given subscriber base is not necessarily disclosed by ARPU figures. This ARPU limitation necessitates the consideration of more indicative metrics of profitability and business performance, which should be more closely related to telecom services that are consumed rather than to revenues that are generated.

ACPU metrics

Average connectivity per user (ACPU) is a new metric, proposed to assess the business opportunities and/or performance of a telecom market. ACPU is calculated as the total type of generated traffic (voice/data) divided by the number of subscribers during a predefined time period – typically a month. Voice ACPU (ACPU-V) is calculated as the number of voice minutes consumed by a subscriber, while data ACPU (ACPU-D) is calculated as the number of megabytes consumed by a subscriber in a given period. In either case, the total considered traffic (voice/data) includes on-net and off-net traffic, incoming and outgoing traffic, and national and international traffic. More interestingly, these new metrics could be tuned to on-net ACPU, off-net ACPU, international ACPU, roaming ACPU, for data as well as voice services.

Broadly speaking, if ARPU focuses primarily on operators’ businesses, ACPU’s appeal is widened to more players, including operators, vendors (networks designers/optimisers), content developers, and so on. In other words, ACPU better reflects the attractiveness and the potential of a given market. ACPU may also better reflect new usage trends such as social networks, voice-to-data migration (given the popularity of smart phones), and the decline of international voice traffic in the face of IP connectivity and PC-to-PC calls. On another hand, ARPU indicates essentially the financial performance of an operator or the financial opportunities of a market, while ACPU addresses the maturity of the telecom market in a country, the volume of traffic (data/voice) generated, and consequently, the willingness of subscribers to consume telecom services.

From another perspective, many telecom operators express dismay at declining ARPUs and consider investing and serving low-income markets as risky. As such attempting to generate a profit from serving low ARPU customers is viewed as being a difficult thing to achieve. However, business logic and the dynamics of the telecom industry are showing this not to be the case, as ARPU figures are financially correlated to operational costs, labour costs, and average income. In other words, high-ARPU climates are not necessarily drivers of profitability, particularly if operational costs are too high and/or not fully allocated. Moreover, if ARPU is used to indicate how successful an operator had been in inciting users to spend more on telecom services, that is no longer the case in saturated and/or highly competitive markets, particularly where promotional offers, free services, or free subscriptions are involved.

Such practices dilute ARPU, and thus high ARPU in a saturated market may not necessarily be more appealing than lower ARPU in a market that has average or low subscriber penetration. Thus non-financially based metrics such as ACPU are revealed to be more suitable in predicting the business success in emerging markets. Typical examples could be taken from the Asia Pacific region, where monthly ARPU ranges from less than US$3 in countries like Bangladesh and Sri Lanka to more than US$50 in Japan.

While ARPU continues to be a widely used metric by telcos, it is time to raise awareness for more realistic and reflective metrics as mobile markets hit saturation and subscriber acquisitions slow down. I also believe it is time to contest assumptions such as low-ARPU markets are unprofitable, while high-ARPU ones are. It is suggested that ACPU metrics better track voice and data usage and coupled with ARPU, help to estimate potential revenues. These new metrics are believed to be a better guide for all telecom stakeholders, in order to help them in deploying more profitable services. More importantly, these new metrics will lead to the development of new business thinking (with respect to business strategies, technology deployments, proposed services portfolios, CAPEX and OPEX investments, and so on) about participating within prepaid segments, low income customers, and declining/low ARPU markets.

Mohamed Jamoussi is a senior advisor for Saudi Telecom (STC)