Getting online is the easy part
What 1,500 refugee businesses in Jordan taught us about digitisation
There is a number in the World Bank's baseline survey of Jordanian and Syrian businesses that I still think about.
Thirty percent of Syrian-owned businesses in the sample had launched a new product in the previous year. Among Jordanian male-owned businesses, the figure was eighteen percent.
World Bank baseline survey of approximately 2,100 businesses in Jordan, 2018–2019. Jordanian female-owned businesses, not shown: 13%.
The refugees were more innovative than their hosts.
This runs against the assumption underneath most enterprise development programming — that vulnerable entrepreneurs need to be taught how to build a better business. The Syrian entrepreneurs in that survey were not short of ideas. Many had run successful businesses in Syria and rebuilt from nothing after arriving. What they had lost was not capability. It was everything that turns capability into revenue: suppliers, customers, and a route to market.
The average Jordanian male-owned business in the survey had over seven suppliers and roughly 1,160 clients. Syrian male-owned businesses had fewer than three suppliers and around 40% as many customers. Average annual revenue: JD 117,000 against JD 14,000. For Syrian women running businesses from home, JD 1,400.
Drawn to scale against JD 117,000. World Bank baseline survey, 2017 revenues. Jordanian female-owned businesses, not shown: JD 8,000.
Same skills. Same effort. Radically different access.
So the intervention was designed around access — build the platform, connect them to buyers, and let the products do the rest.
That was the part that turned out to be easy.
The four barriers nobody puts in a proposal
Before a single sale can happen, an informal business in Jordan runs into obstacles that don't appear in any digitisation strategy document. Each one is mundane. Together they are decisive.
Four barriers, four workaroundsThe networking worked. The venue excluded exactly the people the programme existed to reach.
And then nothing sold
With payments solved, shipping solved, language solved and a platform live with roughly 1,500 businesses on it, we started running marketing campaigns.
The customer acquisition cost came back so high the products could not compete.
Not because they were badly made. The craftsmanship was real — Jordan has, by conservative estimates, more than 300,000 Syrian refugees with developed craft skills. The problem was subtler and more uncomfortable: the products were designed for a market that no longer existed, and were now being sold into one nobody had studied.
A hand-crocheted toy made for a Damascus neighbourhood in 2010 is not the same product as one competing for a buyer scrolling a marketplace in Riyadh in 2019. The colours are different. The packaging expectations are different. The photograph has to do work it never had to do across a counter.
So the programme added something that was not in the original design: product design support for roughly 400 businesses.
Worked examples from the programme's design deck, as reproduced in the World Bank report.
Two hundred received design thinking workshops — understanding customers, prototyping, testing. Fifty received one-to-one mentorship, where artisans brought products, took feedback, and came back with revised versions. A hundred and fifty received design guides and video lectures built for their specific category: prayer rugs, preserved foods, women's accessories, stuffed crochet toys, traditional bags, scented candles, game boards, jewellery boxes.
Blocked: consent unconfirmed, faces visible
Blocked: consent unconfirmed, check contract terms
Blocked: two people at frame edge
World Bank report § IV. Slots read from media-manifest.json — rendered only where approved_for_web is true.
That is the intervention that moved the numbers.
What the results actually say
By February 2020, when COVID-19 stopped the project:
- 1,500businesses hosted on the platform
- JD 217,000in business-to-business price offers sent through it
- 6,214customer clicks across 18 e-commerce listings
- US$20,000in actual sales of redesigned products
- US$200,000in potential demand generated from the GCC market
The honest reading of those numbers is the useful one. Of the 70 producers in the design cohort, 30 made sales. That is not a triumph, and treating it as one would be worse than useless to anyone planning a similar programme.
What it demonstrates is a sequence: access alone produced listings. Access plus redesign produced transactions. Every stage before the design intervention was necessary, and none of it was sufficient.
What this means if you are designing a programme
Count the wrong things and you will report success while nothing happens. Businesses onboarded is an input. Listings created is an input. Sales, repeat orders, and businesses still trading twelve months later are results. The gap between those two lists is where most programme evaluations quietly live.
Assume the boring barriers are the binding ones. Bank accounts, courier contracts, interface language. None of these appear in a theory of change. All of them stop a transaction dead.
Budget for market fit, not just market access. This is the finding I would carry into any similar programme. If you are helping producers reach a new market, you are implicitly asking them to compete in it — and the product that succeeded in the old market has no automatic claim on the new one. Design support was added late here because nobody predicted needing it. It should have been in the plan from the start.
Design the exit at the beginning. The Jedad team published a full open toolkit, including code, so that any organisation could build a comparable platform. The funding ended in 2020. The method didn't.
The part we don't get to write about
In early 2020 this was starting to compound. Restaurants were requesting B2B orders of custom tableware. Dubai Image had sent price quotations and sample requests for the corporate gift market. Mercy Corps Jordan and UNICEF were sourcing artisanal products for institutional gifting.
Then the pandemic closed everything, and none of it completed.
I include that because the version of this story where everything works is not a story anyone can learn from. The mechanism was sound. The timing was not. And the toolkit is public precisely so the next team doesn't have to start where we started.
Saleem Najjar is Co-Founder of Umbrella500 and a co-author ofJedad: Creating Market Opportunities for Refugee and Host Community Businesses in Jordan, published by the World Bank with Aminur Rahman, Giacomo De Giorgi and Emad Jaghasi.