African Independence · White Paper · Version 1.1
Launch and Abandon
What happens to African state tourism and heritage websites after the ribbon is cut
Abstract. State-funded tourism and heritage websites across Africa are commissioned as projects rather than maintained as infrastructure. They are budgeted for a launch, not for a lifetime. This paper documents, with dates and observable evidence, what happens afterwards: expired certificates, dead databases, abandoned domains, and in at least one case a government domain serving foreign gambling content to its own visitors.
It sets out an illustrative cost model showing why maintenance is the cheaper half of the equation, records seven case files, lists the questions that remain unanswered, and proposes four changes to how such projects are funded. Every factual claim here is dated and was checked on the date given. Where the author does not know something, the paper says so.
Contents
This paper covers publicly reachable websites operated by, or historically operated on behalf of, African state tourism and heritage institutions. It is written by an independent publisher who has run travel and heritage websites on Ghana and Africa since December 2000, and who uses these state sites as a working source almost every day.
Method: each site named below was opened and recorded on the date given. Where a site behaved differently for different visitors, that is stated explicitly. Where a claim rests on a single observation, it is labelled as such. No conclusions are drawn about the motives of any person or institution; where behaviour is unexplained, this paper asks a question rather than supplying an answer.
This is not an audit. The author has no access to contracts, budgets, invoices or hosting records. Any figure described here as an estimate is an estimate, and is labelled as one.
The lifecycle is consistent enough to describe as a sequence.
Funding is granted for a defined project, usually covering design, content production and a launch event. The site is inaugurated, with a minister, a photograph and a press release. The disbursement ends. From that day, nobody holds a budget line for hosting, certificates, software updates, database maintenance or content.
What follows is not dramatic. It is slow. A plugin goes out of date. A certificate expires and browsers begin warning visitors away. A database connection fails and is never restored. A hosting invoice goes unpaid. The contractor's contact person leaves the ministry, and the person who replaces them never hears the domain mentioned. Two or three years after the ribbon, a traveller looking up visa rules finds an error message — and nobody inside the institution knows.
At that point one of three things happens. The site stays online as a frozen shell. The site goes dark. Or somebody else takes it over.
The following is a worked example, not a finding. The institutional figures are estimates by the author and are not drawn from any contract. They are set out so that the reader can substitute better numbers.
Cost per functional day = (initial capital + total maintenance) ÷ days of continuous functional operation.
Case A — a commissioned state portal (estimated). A design and content contract in the low six figures, no maintenance line after handover, and a functional life of roughly two years before failure. At an assumed build cost of US$150,000 over 730 functional days, that is a little over US$200 per functional day. At US$60,000 it is still around US$80 per day.
Case B — an independent network (the author's own figures). Initial setup in the low hundreds of dollars, hosting and domain renewals in the low hundreds per year, and twenty-five years of continuous operation. Spread over more than 9,000 functional days, this works out below one dollar per day.
The ratio between the two is large under every assumption the author has tried. Readers should nevertheless resist the ratio, because it is not the point.
The point is the denominator. Institutional money buys a short functional life because nothing at all is bought for the day after the launch.
Independent sites survive not because they are clever, but because somebody pays the hosting bill every year and notices when the site breaks. That is the entire secret, and it costs less than the catering at an inauguration.
If any ministry, agency or funder wishes to supply the real procurement and maintenance figures for a named project, they will be published here in full and unedited, and this section will be revised accordingly.
The domain of Guinea's national tourism office, tourisme.gov.gn, behaved differently for different visitors. Automated retrieval returned a maintenance notice written in English — on the website of a French-speaking state — under the site name "Office", with a copyright line dated 2022. At the same moment, a human visitor was served an Indonesian-language online lottery and gambling portal, whose logo linked back to the tourism office homepage. Timestamped screenshots are held by the author.
Serving one version of a page to automated retrieval and another to human visitors is a recognised technique for exploiting the search ranking that official domains enjoy. Open directory listings of the site's plugin folders were visible in search results at the time — a common symptom of an outdated installation left unattended.
On a follow-up check the same week, the page covering entry formalities returned a server error, and the error handler itself failed.
Not claimed: that any Guinean official knew of this, benefited from it, or was involved in any way. The author has no evidence of that and does not suggest it. A compromised website is, in the first instance, evidence of neglect — not of complicity.
Working: the visa application portal at paf.gov.gn/visa; the investment agency's job listings; the Simandou 2040 mining programme site.
Online but abandoned: the Service Public visa page, with most onward links dead; the investment agency's "Vivre en Guinée" section, last updated in 2024; the central bank site, carrying content from 2025.
Not reachable or failing: the border police information pages, returning a database connection error; the tourism office; the national airport website; a national tourism office domain; the government portal guinee.gn; and the Ministry of Commerce, Industry and SMEs, which produced first a browser security warning and then an unavailability notice.
The observation that matters is not the length of the list. It is the shape of it. The one page that takes payments works. The pages that explain the country do not. The ministry site in the best condition is the one attached to an iron ore project.
The Musée Virtuel des Arts et Traditions du Gabon opened at UNESCO in Paris on 30 November 2006: five rooms, more than 250 objects, fifteen previously unpublished ethnographic films, and a filmed guide, built in nine months and drawing over 200,000 visitors in its first year. It was built in Adobe Flash, which lost browser support at the end of 2020. Its patron, President Omar Bongo Ondimba, died in June 2009. Every expansion announced for the years 2008 to 2010 — a media library, a Bantu civilisations pavilion, a national parks pavilion — was never built.
Nothing of it is playable today. The Internet Archive holds pages from the site, but a Flash application is a shell that loads further files, and those were never captured. The author recovered 34 screen captures, which are, as far as can be established, the most complete surviving public record of the project. They are published in full at A Museum Made of Light.
The lesson here is about format, not about Gabon. Public money bought a proprietary environment with no migration plan and no exported dataset. The spectacle was funded. The archive was not.
In July 2026 the official website of the Ghana Museums and Monuments Board was found redirecting visitors to gambling and adult content. The site was subsequently cleaned up and has been back online for some time.
This case is included deliberately, because it is the counter-example that makes the whole argument. A compromised government website can be recovered, and recovered quickly, once someone accepts that it is their responsibility.
The problem is not technical difficulty. It is ownership.
touringghana.com has for many years been among the most visible tourism portals for Ghana in search results, and carries historic authority from its long presence online. As of the check date the site is live, runs on WordPress, and its homepage carries a modification date of August 2024. It embeds an automated African tourism news feed carrying items published on the day of the check.
The author has previously observed search traffic relating to Ghanaian state heritage properties arriving at this domain, and considers the relationship between the state's own heritage domains and this site to be unclear.
What this paper does not claim. It does not claim the site is unmaintained: the evidence above shows otherwise. It does not claim to know who owns or operates it. It does not claim that any advertising revenue is generated there, nor that any public official receives any. The author has no evidence for any of that.
An earlier draft of this paper stated a stronger version of this section. The check did not support it, and it was removed. That is recorded here rather than quietly deleted, because a paper about accountability should show its own corrections.
What this paper asks. These questions are put openly to the Ghana Tourism Authority, the Ministry of Tourism, Arts and Culture, and the Ghana Museums and Monuments Board:
Answers received will be published on this page in full and without editing.
Broken websites are usually described as a technical problem. They are better understood as one symptom of unanswered channels generally. In the course of this work the author sent written enquiries to several embassies and tourism bodies. Some replied within days, and at senior level. Others did not reply at all; in one case an arranged appointment at an embassy in Berlin did not take place, and telephone lines went unanswered.
This is one publisher's experience and is offered as such, not as a measurement. But it points at the same underlying condition: a published contact point is only as good as the person whose job it is to answer it.
Beyond the Return is Ghana's flagship diaspora and tourism programme, announced by President Akufo-Addo in December 2019 as the Year of Return drew to a close, and framed as a ten-year project under the theme "A Decade of African Renaissance, 2020–2030". A twenty-member steering committee was inaugurated in May 2020, seven pillars were unveiled that June, and as recently as April 2024 the responsible minister received the steering committee and stated publicly that the initiative had come to stay.
Its website, beyondthereturngh.com, no longer works. As of the check date the domain returns Cloudflare error 522: the domain still resolves, Cloudflare's network still answers, and the origin server behind it does not respond at all. The Internet Archive holds 143 captures of the site between 20 September 2020 and 10 July 2026, with error captures appearing from June 2026 onward.
Two details in this case are worth more than the failure itself.
Somebody is still paying for the domain. A 522 is not an expired registration and not a deleted domain. Registry records show beyondthereturngh.com created on 12 September 2019 — three months before the campaign was announced — with the record updated on 10 September 2026 and registration now running to 12 September 2027. The name servers still point to Cloudflare, and the domain carries the standard transfer lock, which means an account holder with access set it that way.
Read those dates together. The website stopped responding in June 2026. The domain was renewed in September 2026, three months later, for a further year. Somebody with access to the account paid to keep the name alive while the site behind it served nothing but an error to every visitor.
That is not a case of a project cut off without warning. It is the signature of a registration running on autopilot inside somebody's account, while the hosting that made it a website had already lapsed and nobody checked.
The search engines have not noticed. Months after the site stopped responding, fragments of its text were still being returned by search engines and by automated retrieval, served from cached copies. A site can be dead to every visitor and still look alive in search results for a long time afterwards. This is one reason such failures go unreported: the people responsible are not told, because the internet keeps showing them what used to be there.
In fairness: parts of the campaign's content survive elsewhere, on the Ghana Tourism Authority's visitghana.com and on the tourism ministry's own site, and a change of government in January 2025 will have altered the priorities attached to a predecessor administration's initiative. Neither fact answers the question this case raises.
The programme runs to 2030. The website stopped in 2026. Nobody announced it, because announcements are for openings.
Not claimed: that the campaign itself has ended, that any funds were misused, or that any individual is responsible. What is recorded here is a dated technical observation and the gap between a stated ten-year commitment and the state of its public front door in year six.
Funding structures reward what can be photographed. An inauguration produces a minister at a lectern, a press release and a deliverable that can be reported to a board. Five years of certificate renewals, database backups and security patches produce nothing anyone can cut a ribbon in front of.
Three consequences follow, and all three are visible in the cases above.
Proprietary formats get chosen, because they demonstrate better on launch day — and then cannot be archived or migrated once the technology moves on.
No named office is accountable after handover. The contractor's obligation ends at final payment; nobody inside the institution inherits it.
Nobody is watching. In every failure described here, it was discovered by an outsider rather than by the owner of the domain — and in one case the search engines went on displaying a dead site for months, which is precisely how such failures stay invisible to the people responsible for them.
Fund the decade, not the launch. Development finance for a public digital project should put a minority of the capital into the build and hold the remainder for hosting, security and content maintenance across a defined period of at least ten years, released annually against evidence that the site is actually functioning.
Mandate open standards and an exportable archive. No public money for formats that cannot be migrated. Every funded project should deliver, as a condition of final payment, a complete export of its content and data in open formats, deposited with at least one institution outside the commissioning ministry.
Name a responsible office, publicly. Every state website should publish the office — not the individual — accountable for its operation, together with a working address for reporting a fault. Every funded project should be handed over to a named post, not to a filing cabinet.
Audit the domain estate. Ministries should maintain and publish a register of every domain they own or have ever funded, including expired and redirected ones, with the hosting renewal date for each. Domains no longer in use should be retired deliberately and kept out of resale rather than abandoned. Commercial advertising on domains funded from public money should be disclosed and, where the domain remains state-linked, prohibited.
Every dated observation in this paper was made by the author on the date stated, and supporting screenshots are retained. Sites that behaved differently for different visitors are described as such, with both behaviours recorded.
This paper makes no allegation of wrongdoing against any named or identifiable individual. Where the author considered a stronger interpretation and could not support it with evidence, the paper says so rather than omitting it. Section 4.5 is the example, and it is left visible on purpose.
Any institution, company or individual named here may respond. Corrections will be published promptly and marked as corrections, with the original text kept visible alongside. Responses will be published in full and unedited.
Version 1.2 — published 22 September 2026, revised 22 September 2026.
Version 1.1 added case file 4.7 on Beyond the Return; version 1.2 added the registry record for that domain.. This paper will be revised as further checks are made, and every revision will carry its own date. Written by Remo Kurka, independent publisher, operating Ghana and Africa travel and heritage websites since December 2000. Corrections, responses and right of reply:
info (at) africanindependence (dot) com