Digital platforms

Measuring the number of digital platform workers is surprisingly difficult

Rory O'Farrell03/05/2019

In the mid-2010s, British researchers surveyed the number of digital platform workers (such as workers using Uber, Deliveroo, or Amazon Turk) by simply asking people whether they earned money on a digital platform. However, they were surprised at the high numbers the results showed. After asking follow-up questions it emerged many had included money won on gambling websites as money earned via an online platform. Clearly, asking a simple question can be a difficult task.

There are three main difficulties in measuring platform workers.

Firstly, there are difficulties in explaining to survey respondents what interviewers mean by platform work. These difficulties are visible in a survey undertaken by the US Bureau of Labour Statistics. They included a detailed description of platform work in their questionnaire. However, many respondents poorly understood the definition, for example in answering yes if they merely made use of a computer or mobile app in their job. After removing obviously incorrect responses (such as hairstylists that said they did their work entirely online), the estimated number of platform workers fell from 3.3% to 1%.

Second, there are inconsistencies across countries in how platform workers are measured. For example, some surveys do not differentiate between capital platforms (such as Air BnB) and labour platforms.

Third, most estimates of the number of platform workers are in the range of 0.5% to 2%. This leads to small sample sizes, limiting how precisely we can estimate the characteristics, such as gender or occupation, of platform workers.

The emergence of digital platform work has led to controversy, such as in August 2018 when New York City banned new ride hailing licences for a year. Platforms facilitate flexible work arrangements, which may lead to an increase in low quality jobs, with poor career prospects. However, it has also been argued that such work has the potential to boost employment, increase flexibility for workers, and to serve as a transition to regular employment. As the current lack of data hinders the ability to give policy advice, there is a need to gather more information regarding platform workers. In particular, there is a need for data that is comparable across countries, across time, and with existing labour market statistics.

The lack of an agreed definition as to who is a digital platform worker hinders their measurement. Platform workers use an app or a website to match with customers in order to provide a service (rather than goods) in return for money. In addition, the use of an app forms an integral part of providing the service, for instance allowing the worker to submit their work (such as product descriptions) through the platform, or making payments through the platform. However, they offer a diverse range of services including ride hailing, coding, and writing product descriptions. Platform work may be a worker’s main job, or occasional secondary work to supplement a worker’s income. Such workers can also vary in the amount of capital (such as a car) they use to carry out their work.

Potential  next  steps  to measure platform workers could  include  collaborative  work among official statistical agencies  to  formulate  standard  questions  for  inclusion  in  surveys. Rather than give survey respondents a detailed definition of platform work, interviewers can ask a series of short questions concerning different aspects of the work performed, and the interviewer can then determine whether the respondent is a platform worker. More experimentation in terms of ordering of questions and use of prompting questions may be necessary before such questions are included in official surveys. 

In addition, alternative data sources may be useful. At present, the possibilities of using administrative data (such as tax records) are limited, but they may increase as tax authorities develop data-sharing agreements with platforms. The use of some existing big-data sources, such as bank records, can also allow researchers to retrospectively refine their research question as new platforms enter the market,  and overcome the problems associated with small sample sizes.

Although measuring digital platform workers is challenging, they are likely to represent a growing share of the workforce. It is time to take on the difficult task of asking a simple question.

More information can be found in the report ‘Measuring platform mediated workers’ available at . The opinions expressed in this article are those of the author and do not necessarily reflect those of the OECD or its member countries.

Posted in: Labour market

Tagged with: cheaplabourprecariouswork

Dr Rory O'Farrell     @r_o_farrell

O'Farrell, Rory

Rory O’Farrell is an economist at the Dept of Economics, OECD, with extensive post-PhD international work experience. His research has been in the area of public policy, labour economics, fiscal policy and monetary policy. He has experience with macroeconomic modelling, calibrating Mortensen-Pissarides matching models and forecasting.





Robert Sweeney

Robert Sweeney is a policy analyst at TASC and focuses on issues surrounding Irish …

James Stewart

Dr James Stewart is Adjunct Associate Professor at Trinity College Dublin. His research …

Rory Hearne

Rory Hearne is a postdoctoral researcher in the Maynooth University Social Sciences …