Mobile internet price comparisons – a review of the Digital Fuel Monitor methodology

[caption id="attachment_7942" align="alignleft" width="300"]digital fuel monitor Source: The Express Tribune[/caption]


From time to time, policymakers, government officials and telco executives find their email in-boxes filled with alarmist, and sometimes unsolicited, press releases warning about mobile internet price increases, measured according to a price comparison methodology called the “Digital Fuel Monitor” (DFM). These press releases include an offer to purchase the full version of the very expensive report to get fully acquainted with its results.


The reports conduct price analyses in the mobile telephony market in different countries of the European Union and the OECD, typically concluding that prices are on the rise, especially in those countries where markets have recently consolidated.


In one of its most recent reports, it was claimed that the prices of Drillisch in Germany had risen by 138% since the merger of Telefónica and E-Plus, and O2's prices in that country had increased by double digits in the same period.


Such statements seemed rather surprising in view of the actual evolution of mobile internet prices in Germany and the reported ARPUs of the companies, and thus likely inaccurate. The DFM report had already been challenged in the past both by Vodafone and by Frontier Economics so we were reluctant to spend a large amount of money to buy it in order to understand a set of conclusions that we did not share. Instead, Telefónica decided to dedicate those resources to assess the validity and consistency of the DFM methodology. We approached the consulting firm Solchaga Recio & Associates (SR) to conduct a review of the methodology, which we are glad to make hereby freely available.


The findings are conclusive and reveal a serious conceptual and methodological flaw in the DFM approach. The flaw leads to ungrounded results, nevertheless marketed with no qualms.


DFM’s method is based on the study of a metric called “price per incremental gigabyte” on the grounds that it is the one used to calculate competitiveness in the electricity sector, in this last case using €/kWh.


SR’s review explains that this metric is generated by plotting a range of data packages commercially available on a graph showing the data package size (GByte) versus price (€) from an operator, and then drawing a trend –regression– line. Then, the report uses as a measure of “price”, the slope –being in €/GByte – of the line: the greater the slope, the greater the price.


To assess price changes –either at different points in time or prices by different providers– DFM quantifies the difference between the two lines’ slopes. SR’s analysis shows that what is actually being measured is the difference in unit prices between the two group of data packages being compared, not a change in the overall average price, which is a very different thing from the price per unit of the packages themselves. It would be the same as saying the distance travelled by a car is its speed, whereas the speed is the distance travelled over a given time – or the slope of the line plotting distance versus time.


This conceptual mistake is explored by SR using crystal clear examples. SR moves the points on the line –data package size or price– in ways that are beneficial to customers: either by reducing the price per unit increasing the data allowance volume of data provided, or by reducing the package price. Then, they assess the results yielded by DFM’s methodology.


The review concludes that according to such methodology, the launch of a new data package plan with a larger GByte allowance and higher price, could be understood as if the carrier has undertaken a real price increase, even if the new plan would actually have the lowest price per unit than any pre-existing plan in a given market. There are cases where even a reduction in the price of a plan is also interpreted as a price increase as it increases the slope of the regression line.


Lastly, SR’s review also shows how the DFM methodology would lead to the conclusion that operators with lower prices are more expensive than others whose prices are actually higher.


The nonsensical nature of these results brings the flaws of the analyzed methodology into sharp focus. We are happy to have spent our money gaining an understanding of the DFM methodology and its shortcomings.  We suspect this will lead policymakers and executives to treat the figures advertising the DFM’s findings with the utmost precaution.


You can –at zero cost–- read the full report here.

Add Comment
Thanks for your comments,

We read them with interest but believe that unfortunately they do not address the key issue and reason why we have published this post: The serious flaws of the DFMonitor methodology means it is not suitable to make price comparisons, as was shown in detail by Solchaga Recio (SR) in its report.

You recognise in your comment that “price per incremental gigabyte alone clearly cannot fully characterize price level differences between markets and operators”, but unfortunately that is exactly the measure you used when you claimed the price increases in the German market that caused us to ask SR for advice and clarification.

To be more specific: You keep claiming in your comment “the German weighted average gigabyte price was €5.09 (…)” and “So the German gigabyte prices were 2.5 times higher than the UK, almost five times higher than the Danish and 30 times higher than the Finnish gigabyte prices.” After reading SR's report, it is clear that those numbers are based on a flawed methodology which means that they simply are not suitable to describe the situation on this competitive market. The flaws in the DFMonitor methodology mean that your claim that Gigabyte prices were higher in Germany than in other European countries, could actually mean precisely the contrary – they could be lower, as the examples in the report show.

We would have welcomed that DFMonitor acknowledges the general difficulties of measuring prices on markets and also that you review your methodology based on the constructive proposals pointed out by SR. In a press release issued yesterday regarding our post you use strong language and claim that Telefónica "hates" the DFMonitor methodology; that surprised us as we are far from that and actually value this interesting debate. We hope that you understand that it is first of all in your own interest to improve your reports and ensure that they correctly reflect competitive market realities.

Finally, to clarify, in the post we refer to the fact that the results of DFMonitor reports have been challenged by Vodafone and Frontier Economics. However, as you should be aware, the specific DFMonitor reports criticized by them have not used the flawed methodology of incremental Gigabyte that SR analysed and which shortcomings we felt a need to highlight.

Thanks again for your comment and also this interesting debate which shows the difficult task of measuring prices in competitive and volatile mobile markets.

Enrique Medina
Posted on 14/10/15 15:15.
No error on our side. The Digital Fuel Monitor methodology is sound. Several EU operator groups (among them ETNO member) are using DFMonitor's price per incremental gigabyte in their investor relations material as the main measure to characterise competitiveness of their own markets.

The price per incremental gigabyte is the key competitiveness metric for open mobile internet access. It will become increasingly centric to the EU-wide data roaming and net neutrality (vertical price discrimination, zero rating) policy debate.

Going forward, the price per incremental gigabyte will gain even more importance and exposure for LTE-Advanced and 5G-based open mobile internet access (1000x capacity increase forecast by leading vendors).

The gigabyte price (slope of the linear trend line) is an accurate and the most relevant metric if the base prices (prices for zero gigabytes) of the operators or countries under comparison are the same or roughly similar. Even more so if the base price of the operator or country that has a higher gigabyte price is higher than the base price of the operator or country that has lower gigabyte price. In that case consumers not only pay less for gaining access to smartphone plans with unlimited minutes and SMS but also pay less per gigabyte of use. Finally in some extreme cases (e.g. Cyprus duopoly where the base price is very high (€79.15) while the gigabyte price is moderate (€2.74) an incremental gigabyte price comparison must be complemented with the basket comparison in order to grasp the full magnitude of the base price.

*Germany* post-merger mobile internet access prices went up, also from "basket perspective”. Before the Telefonica Germany – E-Plus merger the lowest price for the 10GB smartphone plan basket was €45. Post-merger the prise rose to €73 (+59%). Before the Telefonica Germany – E-Plus merger the lowest price for 5GB smartphone plan basket was €36. Post-merger prise rose to €43 (+19%). For further context, for some Swedish operators 30% of the "small screen" mobile internet users are now on 20-50-100 GByte tariff plans and a large share on the 5GByte ones -- this should give some guidance on choosing the most relevant baskets for competitiveness assessments in Germany. The per capita mobile internet consumption in Sweden is some 10x higher than that in Germany (no wonder).

Our full response can be read here (incl a 2-pager pdf with a price-GByte scatter plot of actual Telefonica prices in Germany vs some other markets. Watch the slopes!)

We also sent some responses on twitter at
Posted on 14/10/15 15:15.
Some quick reactions:

1) Contrary to the claim of the blog post and the SR report, the title of the Digital Fuel Monitor report referred not to mobile internet "prices" in general but to "gigabyte prices" – Please kindly correct this in both the blog post and the report

2) DFMonitor tracks and analyses not only price per incremental gigabyte (and base prices) but also basket prices (separately for smartphone and dedicated data-only plans). Check for example

3) In our analyses we directly compare gigabyte prices when the base price is equal or higher otherwise we compare basket prices. For example operator A base price is €30 and operator B base price is €20. Operator A gigabyte price is €5 and operator B gigabyte price is €0.5 then we draw the conclusion that the gigabyte price of operator B is ten times less than the gigabyte price of operator A.

4) While the price per incremental gigabyte alone clearly cannot fully characterise price level differences between markets and operators, it is becoming increasingly relevant in the context of mobile video and cloud service zero-rating practices (one hour of SD video consumes about 1GB, one hour of HD video consumes about 3GB, that is the highest post-merger allowance for Yourfone under Drillisch, pre-merger, under E-Plus they sold up to 10Gemoticon

5) Before the Telefonica Germany – E-Plus merger the lowest price for the 10GB smartphone plan basket was €45. Post-merger the prise rose to €73 (+59%). Before the Telefonica Germany – E-Plus merger the lowest price for 5GB smartphone plan basket was €36. Post-merger prise rose to €43 (+19%).

6) To put the German prices and gigabyte allowances into perspective: TeliaSonera sells in Finland a 50 (fifty) 4G (150Mbps) gigabyte smartphone plan with unlimited minutes and SMS for €30 Additional gigabytes cost €0.2. That is 25 times less that the gigabyte price of Telefonica, T-Mobile and Vodafone in Germany! The competition effect – In Sweden, a 4 operator market, 4G smartphone volume caps doubled (for same prices)

7) According to DFM public data (1Q2014) the German weighted average gigabyte price was €5.09 (base price was €33.46), the UK gigabyte price €1.95 (base price was €21.11), the Danish gigabyte price was €1.08 (base price was €35.39) and the Finnish gigabyte price was €0.17 (base price was €18.29). So the German gigabyte prices were 2.5 times higher than the UK, almost five times higher than the Danish and 30 times higher than the Finnish gigabyte prices.

+1) We are not aware that either Vodafone/Communications Chambers or GSMA/Frontier Economics has challenged the Digital Fuel Monitor methodology
Posted on 14/10/15 15:15.
We have been following Telefonica´s Public Policy Blog in the latest days after Enrique Medina published our study reviewing DFM’s incremental gigabyte methodology. We have read with great interests DFM’s posts on the blog and we would like to respond in order to reconfirm the conclusions of our study, react to what has been said in those posts and to clarify the goals and scope of our report.

Please find below the link to our response:
Posted on 14/10/15 15:15.
Thanks for your comment. As we said, an interesting debate in which we seem to disagree. The intention was to make you aware of a major conceptual flaw in your report which should be in your interest. You can follow up directly with Solchaga Recio regarding the methodology.
Posted on 14/10/15 15:15.


Enrique Medina

Chief Policy Officer of Telefónica