One of the in-depth interviews ran a quiet contradiction past us that took a few minutes to register. The founder was describing a software-publishing project the firm was building — a B2B SaaS for the local restaurant sector — and was clear about what made the project his real interest: the company would own the product, set the price, recur the revenue, never have to chase a contract again. The pride was unmistakable. Then, almost as an afterthought, he mentioned the outsourcing work the firm was doing for foreign customers in parallel. «That’s just to keep the lights on while the product matures.»
We had heard the same sentence, in slightly different words, in seven other interviews by that point. The pattern was a sub-sector inside a sub-sector. Software publishing — the part of the Kyrgyz IT market where firms develop their own products and own the resulting code — was being financed, almost everywhere, by the outsourcing work its founders had been told they were trying to leave. Of the firms we surveyed in publishing, 42.5% were simultaneously doing outsourcing work. The split was not a strategic positioning. It was a survival arrangement.
That arrangement is the centre of gravity of the publishing chapter, and worth being honest about up front.
What the count looked like
- Specialists working in the sub-sector: 1,152.
- Number of organisations: 384 firms registered with the Ministry of Justice (publishing-related NACE codes).
- Total market revenue: not determinable from any single source — KG accounting categories don’t isolate software-publishing revenue cleanly.
- Growth rate: 24.5% average annual growth in firm count over a ten-year window. By share of the IT sector in count terms, software publishing was 40.7%. The largest sub-sector by firm count.
- Gender composition: 94% male in development roles; women in publishing were more often in project management.
- Regional distribution: Bishkek 75.8%, Osh 9.9%, Chui oblast 3.4%, Jalal-Abad oblast 3.9%, Naryn 2.3%, Batken 2.1%, Talas 1.3%, Issyk-Kul 0.7%, Osh oblast 0.5%. The most regionally distributed of the IT sub-sectors. Not by much, but visibly more than outsourcing.
- Named players (local product companies): Namba Taxi, Besmart. Top product companies working internationally: Makeuseof.com, Picvipic.com, Socialshopwave. Plus the breakout case the chapter spent a full paragraph on: Behavox, the fintech compliance company founded in London by Erkin Adylov, recognised in 2018 as one of the UK’s fastest-growing businesses, valued at $300 million, included in the Forbes Top 50 Fintech.
Where the growth was coming from
The single chart from this chapter that travelled best, when we presented the report to government and donor audiences, was the comparison of firm-count growth between publishing and outsourcing.
The number that made the cleanest argument was the 24.5% annual growth in publishing firms versus 11.9% in outsourcing. Both numbers are positive; both are credible; but publishing was creating new firms more than twice as fast as outsourcing was. The 61-to-77 firm jump between 2017 and 2018 was the recent slice of that longer trend.
What was driving the growth was, mostly, the platform economy. The chapter quoted Harvard Business Review’s pipelines-to-platforms framing — Alibaba, Amazon, Uber, Airbnb, Baidu — and applied it to what we were seeing in Bishkek. The biggest local product was a taxi service (Namba Taxi). The next biggest was a Telegram-channel-driven classifieds operation (Lalafo, originally Estonian, with Ukrainian engineering roots, dominant in the second-hand goods market). The third was diesel.elcat.kg, the older bulletin board that Lalafo had displaced in the buy/sell segment. The fourth was the lattice of mobile-wallet apps from banks and operators.
None of those companies were thinking of themselves as competing with a Kyrgyz peer. They were competing with WhatsApp, with Yandex Taxi, with Wildberries, with 2GIS. The competitor list, in the chapter, named global platforms first and named local ones only as small inserts.
The two business models
The publishing chapter ran two parallel typologies that publishers had to choose between when planning monetisation:
B2C products and their monetisation. The most successful B2C products are free at the point of use, monetised through advertising — Facebook, WhatsApp, Google in the global examples; ts.kg, oc.kg, namba.kg locally. Within the platform-economy ad markets, four pricing models were the standard: Cost-per-Click (CPC), Cost-per-Mille (CPM, $X per 1,000 impressions), Cost-per-Action (CPA), and Cost-per-Install (CPI). Local platforms additionally monetised by charging users to promote their own listings (lalafo.kg’s listing-boost model). Mobile applications were free, shareware (basic features free, advanced paid), or paid; 31% of free apps were monetised by in-app ad placements through Facebook, Google, RTB House, and Apple Store Affiliates.
B2B products and their monetisation. Subscription-based access with premium tiers had become the dominant model by 2018, following Jira’s $10 per user per month pattern. The chapter named Artwaga.com as a Kyrgyz example providing B2B solutions (R-keeper, Store House, iiKo, Mobi-S, Ban Agent) to local HoReCa firms — cafés and restaurants. The B2B customer base in Kyrgyzstan was: financial institutions, restaurants, cafés, airline agencies, online stores. Their needs: CRM, ERP, marketing automation, process management. The friction the chapter named, again and again: banks were buying expensive foreign solutions, watching them age, then paying again to migrate to the next one.
Payment methods — the part that was changing fastest
Within the publishing sub-sector, 78.9% of payments were non-cash by early 2019; bank transfer (Visa) accounted for 69.2% of those. The 21.1% cash remainder was concentrated at the Instagram-store end of the publisher spectrum — which the e-commerce chapter would cover in more depth. The non-cash share had been growing year on year. By the end of 2020 — writing this with hindsight — the pandemic accelerated that shift in a way the report could not have predicted: by April 2020 the cash share was a fraction of what it had been when the field research closed in March 2019.
What the 240 apps in the local market looked like
Of the 240 applications hosted on Apple Store and Google Play targeting the Kyrgyz market, the category distribution was:
- Maps and navigation: 18% — but 87.5% of that was taxi-service apps (Bishkek 77%, Osh 20%, Naryn 3%). The remaining 12.5% was mobile maps of public transport routes. Effectively, «maps and navigation» was a synonym for «taxi apps».
- Finance and payments: 12% — currency rates (37.9% of this category, named: «Exchange Rates of Kyrgyzstan,» «Course Som: Dollar, Euro, Tenge»), mobile wallets of mobile operators (34.5%: Balance.kg, O! Money), mobile banking (27.5%: ELSOM, Elcard.Mobile, Mbank online, RSK 24, DemirBank).
- Business applications: 12% — narrowly-targeted (68%: Waiter.KG, SBS KG), sales/purchase (20%: BuySell), job search (12%: Ustalar.kg).
- Work, music and audio: 10% each. Work apps tracked courier orders (46%), accounting documents (38%), and specialist search (17%). Music apps were mostly radio-station apps (74%) or compilations (26%).
- Everything else — news, shopping, transport, social, food, entertainment, communications, travel, mobile TV — under 10% of the total.
The downstream effect of just the taxi-app category, the chapter noted, was striking: about 60,000 taxis in the country (58,700+ driver-app downloads), of which 25,000–30,000 were in Bishkek. Around 5,000 active in Bishkek on any given day. These were the non-technical platform jobs the multiplier chapter had been pointing to. The number of jobs in the courier and platform-employment categories, by app-download proxy, was another 10,400+.
The investment landscape
This is where the chapter’s tone shifted slightly, because the data was so sparse. We counted roughly 20 visible private and institutional investors actively supporting 64 IT projects, with a total deployed capital under $1 million USD across all of them. The named investors: The Farm, Highland Private Equity and Mezzanine Fund (HPEMF), BT Investments, the Kyrgyz-Russian Development Fund. Private individuals — business angels — operated outside any formal association. There was no venture fund in the country that could finance a Kyrgyz publisher from growth stage through to strategic exit or IPO. The HPEMF, a direct-investment fund created in partnership with the International Finance Corporation with a $15M corpus, intended to deploy $0.8M to $5M per IT project — a sensible ticket size for the publisher tier the chapter was studying, but only 64 projects in the whole country had received any equity investment of any kind.
Bank lending against intangible assets had not happened. One respondent in 40 had applied for a bank loan to support a publishing project. The acceleration programmes the country had run — Ideagrad, Startup Bootcamp, between 2015 and 2016 — had provided training and mentoring more than capital. By the time of the 2019 research, those programmes were dormant.
The respondent quote the chapter ran on this was longer than most. We are including it in full because it remains, in 2020, the most accurate summary of what investment looked like in Kyrgyzstan when a software publisher tried to raise a round:
«There should be a layer of people who have a lot of money, who are not afraid of losing and they should have a culture, rules of the game: at what stage they give money, how much interest they get. Our people do not know these rules, these rules are more rigid, archaic, the businessman becomes the owner, there is no understanding of the difference between share and debt. If the company goes bankrupt, the share disappears. If you give a loan, you have the right to demand interest and the return of the principal debt. If local businessmen invest, they ask for a key stake, if they become a co-owner and have received a stake, then they still require bankruptcy as a debt. For this to happen, culture, successful precedents, there should an investment appeal for people. They need to see what they can earn. At the moment there were no cases to confirm.»
The single line that names the problem, inside that quote, is the line about no successful precedent. Behavox was based in London. There was no Kyrgyz-founded, Kyrgyz-invested, Kyrgyz-grown publishing company that an angel investor in Bishkek could point at to make the case to their own peer network. The investment culture would not form, the founder was saying, until at least one such case existed.
The problems publishers named
The cross-sector eight applied here in their own configuration. Three were sharper than the others in publishing:
Execution of local-customer contracts. Local Kyrgyz customers — banks, retail chains, restaurant groups — would order a B2B product, accept the delivery, and then renegotiate the price downward, miss payment deadlines, or violate other contract terms. The respondents named this as the most consistent operational friction. Several publishers actively avoided local contracts as a result.
Lack of investment to support startup projects. As above. The funding mechanism for a Kyrgyz publisher was: bootstrap with outsourcing revenue. The chapter listed Qoovee.com as a success-in-progress story — a Kyrgyz-developed B2B e-marketplace that had attracted $300,000 in investment and had 21,500 visitors in February 2019 — but Qoovee was an outlier.
Lack of attractive projects. Some respondents were blunt that the constraint was not just capital but ideas. «The main problem is that there are no live projects. As soon as living projects arise: there is a strong team, the product is clear and the monetization is clear — people are in line.» Other respondents tied the shortage of project ideas to the small size of the country’s economy and the high level of corruption.
Competition in the global market. Some respondents had concluded that building was no longer worth it. «It’s hard to compete in the global market, even now, honestly, like an IT project, we have very big questions in our development. Because with each day, the cost of the program, development are cheaper abroad. The big question is to develop it yourself or buy ready-made solutions that are several times cheaper and can do more. The main reason for this is that our market is very small.»
Shortage of human resources. Especially marketers and senior product people, named by most respondents. One respondent from Osh added: «I think for at least 3 years, it is necessary to make investments in education. … We in Osh now have only a course for an android developer and a web designer.»
Patent and copyright
One of the dimensions the chapter spent more time on than the outsourcing chapter did was intellectual property. To publish a product internationally, the publisher needed a defensible legal claim to the code, the brand, and the design — and the route to that claim, the chapter found, was the Madrid Protocol via WIPO, not the Kyrgyz patent office. One respondent’s quote, which we have heard echoed several times since:
«Certification is not a cheap and durable thing. Next is the patent. Let’s suppose that our software has a local patent, but this patent does not comply. Further, let’s suppose, we applied to the Madrid Protocol. This is an international patenting. We received it recently and it took two years, but now we are sure that no one will take our brand or technology.»
The two-year delay to file via the Madrid Protocol was, by the time of the research, a known cost. The alternative — registering only through Kyrgyzpatent — was not a viable path for international publishers. The recommendations chapter, which we’ll come to later in the series, made the case for Kyrgyzpatent reform on exactly these grounds. That recommendation moved partially in 2020.
What the chapter recommended
The recommendations the chapter put forward to USAID fell into four clusters:
Create a supportive environment for the sub-sector. A «regulatory sandbox» for fintech — companies could test innovative products, services, business models in regulated areas without going through full government approval. Rule-sets for small towns to test pilot solutions on a limited-duration basis (the example given was traffic-violation video surveillance software). A template package of legal and organisational documents and open-source software hosted by the Ministry of Justice to make new-firm registration faster. A clearer normative framework for investor-publisher interactions: share allocation to employees, share delineation, convertible bonds, equity terms. The ability for local companies to receive payments through Paypal, Stripe, Indiegogo, Kickstarter into Kyrgyz domestic bank accounts.
Increase cross-sector interaction with academia. Training centres for international companies — Microsoft Innovation Center, Google Learning Center, IBM Learning Center, Facebook Learning Center — on the basis of existing universities. Government programmes to stimulate cooperation between large businesses and educational institutions. Innovation vouchers — grants for SMEs to receive academic R&D services. Regular trainings on design thinking, unit economics, product analytics, project management, Google Ads, Google Analytics, Firebase, Marketing Platform. Industry-specific acceleration programmes (the example given: a processing industry accelerator).
Increase export potential through consulting and information work focused on helping domestic publishers understand and enter EAEU and EU markets.
Finance. Prepare and adopt a venture-activity law drawing on neighbour-country experience. Stimulate the creation of a business-angels institution through information campaigns and training. Stimulate the private sector to purchase domestic software (preferential loans for business automation; cattle-breeder loans for livestock-chipping registrations were the worked example). International integration of domestic payment systems.
What changed between then and now
By the second half of 2020, three things in this chapter had shifted, two for the better, one not at all.
The regulatory sandbox at the National Bank — under the special-regulatory-regime provision adopted in July 2018 — was operational by mid-2019 and had started accepting applications from local fintech publishers. That recommendation moved.
The Madrid Protocol registration path — while still slow at the Kyrgyz end — had been used more widely by 2020, including by several of the publishers we had interviewed. The KG-Labs report’s recommendation on Kyrgyzpatent reform had not yet been formally acted on, but the workaround (going via WIPO directly) had become standard practice in a way it wasn’t in 2019.
The venture-activity law — drafted, redrafted, reviewed, sent to multiple ministries — had not been adopted by the end of 2020. The «no successful precedent» problem had not been solved either. Behavox in London was now valued at significantly more than $300M USD; no Kyrgyz-resident-founded publisher had reached even a fraction of that.
The next post in the series turns to e-commerce — the sub-sector that publishers’ platform products feed into, and the one the report named first when asked where the women’s jobs are.
Source: USAID Enterprise Competitiveness Project (2019). Analysis of the Information Technology Sector in the Kyrgyz Republic. Implemented by KG Labs Public Foundation; commissioned by USAID ECP / Nathan Associates / ACDI-VOCA, June 2019. Software publishing sub-sector covered in primary research January–March 2019: in-depth interviews with publisher-founders, B2B SaaS operators, investors named in the chapter (HPEMF, KRDF), and KSSDA representatives.