E-commerce for export: Experimental evidence from Tunisia

Last registered on September 19, 2022


Trial Information

General Information

E-commerce for export: Experimental evidence from Tunisia
Initial registration date
September 19, 2022

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
September 19, 2022, 4:19 PM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.



Primary Investigator

Technical University Berlin

Other Primary Investigator(s)

PI Affiliation
Technical University Berlin
PI Affiliation
Mediterranean School of Business
PI Affiliation
Humboldt University Berlin

Additional Trial Information

On going
Start date
End date
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Trade is an important channel for firm growth and economic development. E-Commerce offers new opportunities for SMEs in developing countries to market and sell their products abroad to more, and potentially more distant clients. As a consequence, SMEs may be able to sell at higher prices, generate higher profits and increase their margins at the expense of intermediaries. However, technology adoption often remains low among SMEs in developing countries, due to supply-side related constraints like lack of knowledge, resources or demand-side factors such as lack of clients with ability to pay for higher quality goods. In this randomized controlled trial, we assess whether the provision of training and technical assistance increases the adoption of e-commerce technology among SMEs in Tunisia and also investigate its impact on exports. We further investigate the intervention's impact on more immediate outcomes such as web traffic, social media followers and perceptions about the technology.
External Link(s)

Registration Citation

Bouziri, Amira et al. 2022. "E-commerce for export: Experimental evidence from Tunisia." AEA RCT Registry. September 19. https://doi.org/10.1257/rct.10073-1.0
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details


The intervention is part of a larger export promotion program for SMEs implemented by a development agency in corporation with the Tunisian export promotion agency. The objective of the intervention is to increase the adoption of e-commerce technologies and digital marketing practices by firms to improve their visibility to foreign clients, in particular in Sub-Saharan Africa, with the ultimate goal to increase and diversify the exports of the participating firms. The interventions has three phases:

The first phase consists of 27 hours of classroom training (three 7 hour presence sessions and two online session of 3 hours each) where information about the benefits and technical implementation of different e-commerce channels (website, marketplaces, social media etc.) and specific e-commerce solutions for export are presented. Due to the Covid Pandemic we opted for a hybrid model with both in-presence and online parts. Furthermore, digital marketing practices such as search engine optimisation (SEO), Google Analytics, online advertisement and how to create a digital communication strategy and online content creation are also discussed. Participants are split into 12 groups of 6-11 participants and the training sessions were held in four major Tunisian cities to reduce travel time for participants. Teaching in groups enables peer discussion and exchange, which we believe is important when teaching about digital technology. In addition, teaching in groups enables significant cost-savings and is one promising avenue for export promotion agencies to reach out to more firms. The aim of this first phase is to improve participants' knowledge about the benefits and technical possibilities that different e-commerce solutions and digital marketing practices offer. The goal is to persuade them to invest resources to expand and develop e-commerce and digital marketing activities. At the end of the training, companies are asked to formulate a list of actions that they aim to implement during the second and third phases of the intervention.

During the second phase, companies receive a student from digital marketing or related areas for 10 days (80h) during a two month period to assist the company in non-technical, preparatory exercises that were defined at the end of the classroom training. Tasks are firm-specific but include the benchmarking of their website against competitors, creation of a digital communication plan or keyword identification for SEO. This phase was created because the training literature suggests that hands-on assistance is crucial to help SMEs adopt the general digital marketing concepts learned. For example,\textcite{Dalton.2020} find in a RCT with small retailers in Indonesia that firms only improve business practices if a handbook about best practices is combined with two individual assistance sessions. Secondly, having students assist the firms seems promising in several regards. Senior businessmen and women may be intimidated by experts, in particular when it comes to the use of computer and digital tools. They may be more at ease to admit their weaknesses to young students. \textcite{Iacovone.2021} find encouraging results in a RCT with small scale retailers in Mexico City in which students, supervised by senior consultants, can help small firms substantially change their physical marketing with a positive preliminary impact on sales (15-20 percentage points). The objective of this phase is for the firms to develop in detail the action plans they considers most suitable to pursue after the classroom training and prepares them for technical implementation.

For the third and final phase of the intervention, firms receive a total of 17.5 days of customized technical support from an IT consultant to improve their online presence (10.5 days) and digital marketing tools (7 days). The objective of this phase is to technically implement the solutions that companies identified together with the students during the second phase. Due to differences in the original levels of e-commerce readiness we regard this as a crucial step to improve the sophistication of the firms' online presence and digital marketing practices.
Intervention Start Date
Intervention End Date

Primary Outcomes

Primary Outcomes (end points)
1) Weighted E-commerce adoption index (For questions related to an online channel that the company does not have, it will receive zero). An index of standardised z-scores of a series of question relating to the quality of the firms online presence on any of the three observed channels (website, social media, e-commerce platforms)
2) Digital Marketing Index: An index of standardised z-scores of a series of digital marketing practices
3) Online sales: Binary indicator whether the firms incurred any online sales during the year 2023.
4) Inverse hyperbolic sine of online sales in Tunisian dinar in the past financial year (winsorized at 95th percentile)
5) Average organic online traffic of the past 6 months (March 2023 to September 2023)
6) Average Number of foreign countries of organic online traffic of the past 6 months (March 2023 to September 2023)
7) Average number of social media followers:
8) Average review rating of firms' Facebook page.
9) Inverse hyperbolic sine of export revenues (winsorized at 95th percentile) of the past financial year
10) Number of foreign countries sold to in the past financial year
11) Export at all in the past financial year (binary)
12) Trade-weighted income level of export destinations
13) Trade-weighted geographical distance
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
1) E-commerce/Digital marketing knowledge: Index of standardised z-scores of a series of questions testing the firm's knowledge of e-commerce and digital marketing tools
2) Perception of e-commerce technology: Index of standardised z-scores of a series of questions asking the firm to rank the difficulty of a certain e-commerce/digital marketing practices on a 5-point scale
3) Inverse hyperbolic sine of domestic revenues (winsorized at 95th percentile) in the past financial year
4) Total profits made in the past financial year:
3) Product sales margin in %
4) Unit price of the product or service in Tunisian Dinar
5) Number of intermediaries between firm and the final consumer of the product/service
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
Eligible firms will be randomly allocated to treatment or control group, using a stratified randomisation approach. We use stratification because of our relatively small sample size and the fact that we have large variation in the financial size of the firm (measured in total Revenue) and also in pre-existing levels of the adoption of e-commerce and digital marketing practices (see Table \ref{} for details). The random assignment followed the following procedure:
-We form three groups of high, medium and low scores of the sum of all knowledge and adoption questions of e-commerce and digital marketing.
-We form again three groups of firms that have sell to 3 or more countries, 1-2 markets and of firms that do not export at all.
-We form four groups of high, medium, low and missing export revenues
-We then aggregate the groups formed in Step 2 and 3 into six categories of low/high number of export countries and high/medium and low export revenues.
-Groups of all combinations of Step 1 and Step 4 are being created (high, medium, low e-commerce activities) with the six exports group
- 3 Strata for firms that did not export, missing export values and firms with missing data for all strata variables are being created.
- Finally, in two cases strata were merged because they were too small.
Experimental Design Details
Not available
Randomization Method
stratified randomization using randtreat in STATA (Version 15)
Randomization Unit
Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters
236 firms
Sample size: planned number of observations
236 firms
Sample size (or number of clusters) by treatment arms
117 firms in treatment group
119 firms in control group
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
see Pre-Analysis Plan

Institutional Review Boards (IRBs)

IRB Name
IRB Approval Date
IRB Approval Number
Analysis Plan

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information