Citation-friendly pages
MISSING0 of 25 fetched URLs are best-of, alternatives, comparison, review or pricing pages. Everything published so far is an informational guide.
Your site barely shows up on Google for the searches your audience is running. Here's what's broken, and the first moves to fix it.
All 25 pages we fetched have an H1 and schema, every page has a written meta description, and the blog covers practical topics like choosing a POS app, Google Ads for beginners, AI agents and the UU PDP data law.
Cokroweb is an Indonesian digital agency. It sells AI chatbots for WhatsApp and websites, workflow automation, and website builds (company profiles, landing pages, online stores, plus templates for tour and travel, rental and schools).
It also sells POS, HRIS and ERP systems, and SEO and paid ads. The buyers are UKM and online store owners, tour, travel and rental operators, and schools, villages and government offices.
The site is well built.
None of that is reaching Google yet. Our data shows 0 keywords in the top 100 and an estimated 0 organic visits a month, and that count hasn't moved in 6 months.
With content and technical setup in this shape, the problem is that the domain hasn't earned trust yet. Our index returned no backlink figures for cokroweb.id, so we can't size the link profile, but a site with this much decent content and zero rankings almost always lacks link signals.
Get links from Indonesian business sources and agency directories before you write another blog post.
6 structural issues.
Technically the site is clean. All 25 fetched pages returned 200, none are noindexed, none lack a canonical, there are 0 duplicate titles or metas, schema is present on every sample, and content is fresh (newest lastmod is 1 day old). Blog articles are substantial, well structured and written for a real buyer, and the UU PDP and AI usage data pieces show genuine effort.
The problem is that none of it ranks. Our index found 0 ranked keywords across 69 URLs, including 38 service pages and 17 blog posts. This is a young 2025 domain with unknown link authority that spreads itself across six service lines and sends equal internal weight to every page through 220+ links per page. It also aims its blog at head terms that Google and national publishers already own. Better crawl hygiene will not move it. Concentration will: fewer priority pages, one topical cluster and external proof that real clients exist.
The /layanan/ section holds 38 sitemap URLs covering AI chatbots, websites, POS, HRIS, ERP, SEO and ads, and our index found 0 ranked keywords under that prefix. Every money page is invisible in search, so all inbound demand depends on WhatsApp referrals and paid traffic. The likely mechanism for a domain operating since 2025 is a lack of trust signals: Google has no external evidence that this domain deserves to rank for commercial terms like 'jasa pembuatan website' or 'chatbot WhatsApp'. The weak internal routing described in the other findings makes it worse, because 38 pages split what little authority the domain has.
1. Verify the domain in Google Search Console and submit the sitemap if that has not been done.
2. Use the URL Inspection tool on all 38 /layanan/ URLs and record which are indexed, crawled-not-indexed or discovered-not-indexed.
3. Pick the 5 services with the clearest local demand (for example company profile websites, WhatsApp chatbots and POS systems) as priority pages.
4. Give each priority page a unique title built around the Indonesian buyer query plus a city or 'UMKM' modifier.
5. Add pricing, a real client example from /portofolio/ and an FAQ block to each priority page.
6. Hold back new /layanan/ variants until the priority pages start showing impressions.
All 17 /blog/ URLs have 0 ranked keywords despite the sampled articles being full guides with structured H2s. The topics chosen are the problem. 'Cara pasang iklan Google Ads', 'cara daftar bisnis di Google Maps' and 'prompt ChatGPT untuk caption' are contested by Google's own help docs, national media and established agencies with years of links. A new domain cannot outrank them. The posts are also spread across six unrelated service lines (ads, Maps, HR attendance, POS, UU PDP, AI stats), so Google sees no topical cluster where the site is the obvious authority.
1. Stop publishing broad how-to posts on platforms Google documents itself (Google Ads, Google Maps, ChatGPT).
2. Choose one cluster to build first, ideally WhatsApp AI chatbots for UMKM since it is the lead offer.
3. Publish 8 to 10 long-tail posts in that cluster (cost, comparison, industry use cases such as chatbot for tour and travel bookings).
4. Link every post in the cluster to the matching /layanan/ page with descriptive anchor text.
5. Add links back from that /layanan/ page to each cluster post.
6. Reassess rankings for the cluster in Search Console after 90 days before opening a second cluster.
Each fetched page has between 220 and 270 links, including the refund policy (221) and privacy policy (220). That means a mega-menu or footer links to nearly every URL on a 69-URL site from every page. Internal PageRank gets spread flat across all 38 service pages, legal pages and blog posts, so no priority page receives a stronger signal than any other. Google also discounts sitewide boilerplate links, so the contextual links that should mark a service as important are lost among two hundred identical ones.
1. Cut the global navigation to the six service categories plus About, Portfolio, Blog and Contact.
2. Move the links to individual /layanan/ sub-pages onto their category hub pages.
3. Remove sitewide footer links to individual blog posts.
4. Add 3 to 5 in-body contextual links from each blog post to its relevant service page.
5. Re-crawl after the change and confirm legal pages carry under 60 links.
The section hubs that should rank for category terms use bare labels. '/layanan' is titled 'Layanan | Cokroweb' (18 characters), '/blog' is 'Insight | Cokroweb' (18), '/portofolio' is 'Portofolio | Cokroweb' (21) and '/kontak' is 'Kontak | Cokroweb' (17). Titles are still among the strongest on-page relevance signals. A hub titled only 'Layanan' tells Google nothing about what is offered or where, so it cannot compete for queries like 'jasa website dan chatbot AI'.
1. Retitle /layanan to name the core offers, for example 'Jasa Website, Chatbot AI WhatsApp & Sistem Kasir | Cokroweb'.
2. Retitle /portofolio to include 'contoh website' and the industries served.
3. Retitle /blog to describe the topic, for example 'Panduan AI & Website untuk UMKM | Cokroweb'.
4. Keep each new title between 45 and 60 characters.
5. Rewrite the H1 on /layanan ('Enam layanan, mulai dari yang paling mendesak') to include at least one service keyword.
The business has operated since 2025 and our index returned no backlink figures for the domain, so link authority is unknown. Its strongest proof lives in /portofolio/ (8 URLs), yet that page mixes real client projects with 'demo template per industri', which dilutes the evidence that real businesses use the work. Team profile pages exist but are not in the sitemap. For a service business selling to UMKM and government bodies, Google's quality systems look for verifiable experience, and right now the site offers little that a third party confirms.
1. Separate real client case studies from industry demo templates under distinct URLs.
2. Add client name, industry, city and a measurable result to every real case study.
3. List the business on Google Business Profile and Indonesian business directories with consistent name, address and phone.
4. Ask each completed client to add a 'dibuat oleh Cokroweb' footer credit linking back.
5. Add author bylines linking blog posts to the /tentang/tim/ profile pages.
Our crawl found 4 live pages missing from the sitemap, including /tentang/tim/muhammad-ilham-wicaksono, which is linked from 19 pages. Only 23 of 69 sitemap entries carry a lastmod. On a new domain with low crawl priority, an incomplete sitemap slows discovery and gives Google no freshness signal for two thirds of the URLs. Image alt coverage sits at 53%, which also leaves image search and accessibility signals unused.
1. Add all /tentang/tim/ profile pages to the sitemap.
2. Emit an accurate lastmod for every sitemap entry based on real content edits.
3. Resubmit the sitemap in Search Console.
4. Write descriptive alt text for every content image, prioritising /layanan/ and /portofolio/ pages.
3 pages analysed for AI-writing patterns.
Scored 45/100 across 3 pages. The writing has some shape but several tells per page (13 flagged) suggest AI assistance without an editor's pass. Cleanup is straightforward, not a rewrite.
“keberadaan lokasi di peta digital bukan lagi sekadar pelengkap informasi di internet”
The 'not merely X' framing is stock LLM scaffolding that inflates importance, a variant of 'not just X but Y'.
“Oleh karena itu, mematuhi pedoman penamaan sejak awal adalah langkah terbaik untuk menjaga kelangsungan profil bisnis Anda.”
Formulaic summarizing closer that restates the section without adding information.
“Selain mendatangkan kunjungan fisik ke gerai, profil peta digital berfungsi sebagai sarana membangun reputasi usaha yang terpercaya.”
Additive paragraph opener ('Selain...') in a mechanical benefit-listing pattern.
“profil peta digital berfungsi sebagai sarana membangun reputasi usaha yang terpercaya”
Repeats 'profil peta' and 'peta digital' in near-identical phrasings across adjacent paragraphs instead of using pronouns or synonyms.
“Pilar pertama adalah relevansi, yaitu seberapa akurat profil dan penawaran Anda menjawab kueri”
Rule-of-three 'pillars' structure with templated definitions reads as formulaic LLM outline.
“Panduan lengkap cara daftar bisnis di Google Maps lewat Google Profil Bisnis, langkah verifikasi lokasi, dan strategi optimasi”
Keyword-stuffed rule-of-three meta description in generic SEO template style.
“Iklan otomatis memperbesar apa yang sudah ada, baik kelebihan maupun kekurangannya.”
Neat aphoristic closer in a tidy summary, a mild LLM-style flourish, though it is substantive.
“Siapkan beberapa variasi foto, video, dan teks yang menonjolkan sudut berbeda, misalnya harga, kualitas, atau kemudahan.”
Rule-of-three list pattern, common in LLM prose.
“Dulu, beriklan di Facebook atau Google berarti memilih sendiri umur, minat, kata kunci, dan lokasi iklan.”
Formulaic 'before vs now' framing opener, a generic explainer scaffold.
“Kebocoran bisa terjadi bukan hanya karena peretasan, tetapi juga karena akun dipakai bersama”
Classic 'not only X but also Y' parallel construction, followed by a list of three causes.
“Bagi bisnis kecil, risiko terbesar sering bukan denda, melainkan hilangnya kepercayaan.”
'Not X, but Y' contrast used to add a punchy but generic insight.
“perlindungan data perlu dirancang sejak awal, bukan ditambahkan belakangan.”
Another formulaic 'not X but Y' contrast in a promotional closer, typical of LLM-polished copy.
“Anda tidak perlu menyusun dokumen setebal buku untuk memulai.”
Reassurance filler with a templated 'you don't need X to start' shape; list-heavy, FAQ-formulaic structure.
Under 30 characters. Likely missing keyword context.
Under 120 characters. Wasting SERP real estate.
We reached these through your internal links. They load, can be indexed and name themselves as canonical, but the sitemap does not list them, so Google has to find them through links alone.
Fix: Add them to the sitemap, or check whether your sitemap setup skips this section.
Based on the 25 pages that returned HTTP 200 out of 25 fetched. Not a full-site crawl.
Click depth and inlink counts are computed over the 25 pages we fetched, not a full crawl. Pages outside the sample are not represented in the graph.
Compared the first 25 sitemap URLs against the top-500 ranked-keyword URL set.
Classic rankings and AI citations are now separate games. This checks whether the site has pages and entity signals worth retrieving.
cokroweb.id is technically ready to be read: llms.txt is present, all 25 fetched pages have schema, and the homepage meta clearly says AI, websites and Indonesia. It's unlikely to be cited today, though. It doesn't rank in Google, there's no Wikipedia entity, our data shows no off-site or video proof, and none of the fetched pages use the comparison or best-of formats AI answers pull from most.
0 of 25 fetched URLs are best-of, alternatives, comparison, review or pricing pages. Everything published so far is an informational guide.
The H1 "Partner AI & Website untuk Bisnis Anda." is generic. The meta mentions Indonesia, but the title and H1 never name chatbot, WhatsApp or the target buyers.
Schema on 25/25 pages, llms.txt present, and posts like the 2026 Indonesia AI usage data piece cite named sources (APJII, Microsoft, Deloitte, IBM).
No Wikipedia entity was found, our index returned no link figures, and our data shows no video or directory mentions.
With 0 keywords ranked, there are no results pages where an AI Overview could feature the site.
No baseline exists yet. Start with a monthly check of how AI engines answer prompts like 'jasa chatbot WhatsApp Indonesia'.
Publish pages like chatbot WhatsApp vs. admin manual, website template vs. custom, and how to choose a POS for a toko, each linked to its service page. AI answers lean on these formats, and they bring in buyers instead of researchers.
Profiles on agency directories, Crunchbase and Indonesian business listings give AI systems controlled third-party sources that confirm what Cokroweb is. Use /free-backlinks filtered by agency and ai-tools.
Put 'chatbot AI WhatsApp' and 'Indonesia' in the title and H1, and show the client count or portfolio results above the fold. AI engines often cite the homepage, so it should give the answer directly.
One keyword in, one researched article out, every day.
REAL DATA ABOVE · 5 QUICK WINS FOR COKROWEB