Corporate Gifts Singapore: 3 of 8 Sites Show a Price
Contents16
- Key takeaways on corporate gifts Singapore websites
- What we searched, and which sites we kept
- Do these pages say what a gift costs?
- The two questions every gift buyer asks
- How many write their headings the way buyers ask?
- Which of the nine would a machine quote?
- What the structured data declares
- One site in nine keeps an AI crawler out
- Three of nine serve a web page at /llms.txt
- Don't read the speed scores the way they look
- A third of the results are not the suppliers' own pages
- What I would do this week
- What it costs
- Method and limits
- Want these checks run on your page?
- FAQ
An HR executive with a dinner-and-dance in six weeks wants three things from a corporate gift supplier before they send an enquiry: the price per piece, the smallest order accepted, and how long production takes. On 21 October 2026 we read the home pages of nine Singapore corporate gift supplier websites looking for those three answers.
The titles are perfect. The answers are missing. All nine home-page titles contained both "corporate gift" and "Singapore". Of the eight home pages our plain HTTP client could read, three showed a price for the gifts themselves, four gave a minimum order quantity a buyer could act on, and five gave a lead time in days. Two of the eight answered none of the three.
The published lead times tell their own story. Among the five sites that gave a figure, standard production ran from 1 to 3 working days at one end to about 30 working days at the other, and one site quoted 21 to 45 days for bulk runs. Same category, same island, a tenfold spread. The three sites with no figure leave a buyer no way to know where they sit on it.
Key takeaways on corporate gifts Singapore websites
Every figure below comes from one morning's work on 21 October 2026: three searches through SingRank's self-hosted SearXNG metasearch, then our own audit tool and a plain HTTP fetch of each home page. Counts are always shown with their base.
| # | Finding | The number |
|---|---|---|
| 1 | Home-page titles containing both "corporate gift" and "Singapore" | 9 of 9 |
| 2 | Home pages showing a price for the gifts themselves | 3 of 8 |
| 3 | Home pages giving a minimum order quantity a buyer could act on | 4 of 8 |
| 4 | Home pages giving a lead time in days | 5 of 8 |
| 5 | Home pages answering none of those three questions | 2 of 8 |
| 6 | Home pages with at least one heading phrased as a question | 4 of 8 |
| 7 | Opening paragraphs naming the service and Singapore | 7 of 8 |
| 8 | Home pages passing all three on-page checks together | 3 of 8 |
| 9 | Sites declaring a LocalBusiness or Store entity | 5 of 9 |
| 10 | Sites declaring no business entity at all | 2 of 9 |
| 11 | Sites with a real llms.txt file | 5 of 9 |
| 12 | Sites serving an HTML page at /llms.txt | 3 of 9 |
| 13 | Sites blocking one of the five AI crawlers we tested | 1 of 9 |
| 14 | Dead links found across 540 links checked | 1 |
| 15 | On-topic result slots held by pages the supplier does not own | 13 of 36 |
| 16 | Home pages offering WhatsApp contact | 8 of 8 |
If you read one row, read row 5.
What we searched, and which sites we kept
We ran three searches a buyer would actually type: "corporate gifts supplier singapore", "customised corporate gifts singapore" and "corporate gifts wholesale singapore". All three shapes come back from Google Autocomplete for this category, and Google's own explanation of autocomplete says predictions "reflect real searches that have been done on Google", shaped by the location a query comes from. So these are real phrasings, not ones we invented.
From each search we kept the first 12 results, 36 slots in all. Sixteen distinct Singapore supplier domains appeared. We tried to audit 11 of them and finished 9: one never answered our crawler, and one timed out twice.
Each site then got two passes. Our audit tool collected the technical layer. A plain HTTP fetch, with no JavaScript, collected the words, because that is roughly what an AI crawler reads.
One of the nine returned a bot-challenge page to that plain fetch instead of its home page. So page-text counts have a base of 8, and site-level counts a base of 9.
Do these pages say what a gift costs?
Three of the eight readable home pages showed a price for the gifts. The figures were unit prices and entry points: as low as S$0.75 on one product line, from S$3 a unit on another site, from S$1.50 on a third, with individual items listed at S$3.00 and S$4.00.
Two more pages had a dollar figure that was not a gift price. One showed an S$80 surcharge for delivery to an additional location. Another offered a "choose budget" filter with no figure anywhere in the page text.
That gap matters more than it looks. A filter is a tool for somebody already browsing. A buyer comparing four suppliers in four browser tabs is not browsing yet, and neither is a machine summarising the category. Both need a number in the text.
We are not arguing that every supplier should publish a full price list. Custom manufacturing genuinely varies by material, print method and quantity. But "from S$1.50" costs a supplier nothing and answers the question, and our B2B SEO guide for Singapore firms makes the same case for any quote-based business.
The two questions every gift buyer asks
Six of the eight pages mentioned a minimum order quantity. Only four gave a figure a buyer could act on: 50 units on one site, from 50 on another, from 100 on a third, and one supplier stating plainly that it has no minimum at all. The other two offered "low MOQs" and "MOQs vary depending on the product" without ever landing on a number.
Lead time went the same way. Six pages raised the subject; five gave days. One said 1 to 3 working days. One said 5 to 10 business days after proof approval. One said about 30 working days from artwork sign-off. One said 22 working days, or 12 for urgent orders. One said 21 to 45 days for bulk runs.
Read those five figures together and the spread is the finding. A buyer who assumes a fortnight is normal will miss an event by a month with one supplier and pay for speed they did not need with another. Nobody in this sample puts that range in front of them.
Two of the eight pages answered neither question, showed no price and carried no question headings. Those two rank for buyer searches on the strength of a well-written title and nothing behind it.
How many write their headings the way buyers ask?
Four of the eight pages had at least one H2 or H3 phrased as a question. The distribution was lopsided: one site had 51 question headings in a numbered FAQ block, one had six, and two had exactly one each. The other four had none.
Question headings are not a trick. A buyer scanning for "what is your minimum order" finds it, and a system looking for a passage to quote finds a clean pair it can lift without guessing where the answer starts. Our 2026 study of AI search readiness in Singapore found the same signal to be the weakest of three across 109 local pages, so this is not a corporate gifts problem.
One question appears in Google Autocomplete for this category and on none of the eight pages: whether business gifts are tax deductible. Zero of the eight mentioned tax, deductibility or IRAS. We will not state the rule here, because tax treatment is for the tax authority to state and your accountant to apply to your books. But a supplier who answers it in plain words owns a question the whole category has left open.
Which of the nine would a machine quote?
Put three on-page checks together — an opening that names the service and the country, at least one question heading, and a business entity in the structured data — and 3 of the 8 readable pages pass all three.
Here is the part worth sitting with. Three of eight showed a price, and three of eight passed all three on-page checks, but they are not the same three. Only two sites did both. Being readable and being answerable are separate jobs, and most of this category has done neither completely.
What the structured data declares
Five of the nine sites declared a LocalBusiness or Store entity on the home page. Two declared only a generic Organization. Two declared no business entity at all.
Those last two are the interesting ones. Both carried plenty of other markup: breadcrumbs, a website entity, images, a search action, and in one case opening hours and geo-coordinates with no business for them to belong to. The plumbing is in. Nothing is connected to it.
Our guide to schema markup for AI search covers the types. For this sample the lesson is narrow: markup volume is not markup quality, and a page with fourteen schema types and no business entity has told a machine everything except who it is.
One site in nine keeps an AI crawler out
We checked each robots.txt for five AI crawlers: GPTBot, OAI-SearchBot, ChatGPT-User, PerplexityBot and ClaudeBot. Eight of nine blocked none of them. One blocked exactly one, ClaudeBot, and left the other four alone.
That block is worth understanding rather than scolding, because it may well be deliberate and correct. Anthropic's own documentation describes three separate bots: ClaudeBot, which collects content for model training; Claude-User, which retrieves pages when a user asks a question; and Claude-SearchBot, which supports search quality. Blocking ClaudeBot opts a site out of training while leaving the two retrieval bots free to fetch it.
OpenAI's crawler documentation draws the same line. GPTBot crawls content that may train its models; OAI-SearchBot is the bot that surfaces websites in ChatGPT's search features.
So the robots.txt question is not "AI, yes or no". It is which job you are refusing. Our robots.txt census of 219 Singapore sites counts how common each refusal is.
An open robots.txt is also not the same as an open door. The site that blocked ClaudeBot served our plain client a challenge page carrying a noindex instruction, which is roughly what a crawler without a browser engine would get. We measured robots.txt. We did not measure the wall behind it.
Three of nine serve a web page at /llms.txt
Five of the nine sites had a real llms.txt file. Three returned HTTP 200 at /llms.txt and served an HTML page instead of a file, which any checklist counting status codes would score as a pass. One was unreadable behind the challenge page.
The five real files ran from 3.7 KB to 568 KB. The largest is 150 times the smallest, roughly the length of a short novel. One opened by announcing that an SEO plugin had generated it.
A 568 KB file nobody has read is not a signal, and our llms.txt guide for Singapore businesses explains why the file is nowhere near the first fix. If you have one, open it. You may be publishing your entire sitemap as prose.
Don't read the speed scores the way they look
Our tool collected PageSpeed mobile for five of the nine sites. All five scored 71 or below: 57, 57, 58, 69 and 71, with lab LCP from 5.3 seconds to 51.6 seconds. Those numbers look like an emergency, and for two of the five they are not: both had real-user field data showing LCP of 1.7 and 1.2 seconds, inside Google's 2.5-second threshold. The lab test is one throttled simulation; the field data is what visitors actually got. We have written that up at length, because it is the most common way a Singapore owner is talked into a rebuild they do not need: your site scored 16 on PageSpeed, don't rebuild yet.
Across 540 links checked, 60 per site, we found one dead link, an external 4xx. Broken links are not this category's problem.
A third of the results are not the suppliers' own pages
Of the 36 on-topic slots we read, 21 were a supplier's own website. Thirteen were pages the supplier does not own: free blog hosts, a music-platform profile, a document-publishing site, a video channel, two business directories and two third-party blogs. Two were businesses outside Singapore.
Several of those thirteen carried text restating a supplier's marketing copy on a domain the supplier does not control. We name nobody and allege nothing about how those pages came to exist. As a practice it is a poor trade: the page may rank, but you cannot edit it, you cannot put your enquiry form on it, and you cannot take it down.
If a page you do not own holds your category's search, read what it says about you before you spend anything on your own site. Our piece on best-of lists holding 28 of 60 slots works through whether to pay for a slot on one.
What I would do this week
Four edits, in this order, for a corporate gifts supplier in Singapore:
- Put three numbers above the fold. A from-price per unit, a minimum order quantity, and a standard lead time in working days. If one genuinely varies, publish the range and the thing that moves it.
- Turn your six most common enquiry emails into six question headings, with the answer in the first two sentences underneath. You already have the answers. They are in your sent folder.
- Open your own /llms.txt and /robots.txt in a browser. If llms.txt shows your website, you do not have the file. If robots.txt names a bot, know which job you are refusing.
- Search your own category and count the slots you do not own. Then read what those pages say about you.
What it costs
The first two edits cost an afternoon of somebody's attention. They also cost you something real: publishing a from-price and an MOQ loses you the enquiries from buyers who were never going to clear your minimum, and some suppliers would rather have that conversation on the phone. In our view it is a trade worth making, because those enquiries cost a sales team more than they return. But it is a trade, not a free lunch.
The third costs ten minutes. The fourth may cost you a comfortable belief about where your enquiries come from.
If you would rather somebody else ran it, we do: a free visibility check through our contact page, and our SEO, AEO and GEO programme is S$750 a month with 20 articles a month and a three-month minimum. What we will not do is promise a ranking or an AI citation. Assistants choose their own answers.
Method and limits
Read these before quoting any number above.
- Metasearch is not Google. Results came from SingRank's self-hosted SearXNG on 21 October 2026. Eight of its engines returned errors that morning, so the ordering is mostly blended Bing and Yandex, not Google's ranking and not its Maps pack.
- One morning, three phrases. A snapshot of one category on one day.
- Nine sites. Site-level counts have a base of 9, page-text counts a base of 8, PageSpeed a base of 5. Small numbers move a lot on one site.
- Home pages only. We read the home page for the words. A price or MOQ on a product page we did not fetch is not counted.
- Raw HTML. No JavaScript, no login. One site served a bot challenge instead of its page.
- Our audit score is directional. It is our own weighted penalty count, not a Google metric.
- No business is named, and no business SingRank works for appears in the sample. We checked the list before scoring.
Our method page sets out how we separate findings from assumptions.
Want these checks run on your page?
Send us the one page that should win "corporate gifts supplier singapore", or your own category's equivalent. We will run the same checks and tell you which of the three numbers is missing. That is the free visibility check, through the contact page.
These findings describe nine websites sampled on 21 October 2026 through metasearch and a plain HTTP fetch. This page names no business and is not a rating of any company. Nothing here is tax advice.
FAQ
What minimum order quantity do Singapore corporate gift suppliers state?
Among the eight home pages we could read on 21 October 2026, four gave a figure: 50 units, from 50 units, from 100 units, and one supplier stating it has no minimum. Two more mentioned minimums without a number. Two never raised the subject at all.
How long does customised corporate gift production take in Singapore?
The five sites in our sample that published a figure ranged from 1 to 3 working days up to about 30 working days from artwork sign-off, with one quoting 21 to 45 days for bulk runs. Lead time depends on product, print method and quantity, so get yours in writing.
Should a supplier publish prices if every order is customised?
Publish a from-price and the thing that moves it. In our sample only three of eight home pages showed a unit price, so a buyer comparing suppliers can price the rest only by sending an enquiry. A from-price also filters out the orders you could not serve anyway.
Does blocking ClaudeBot stop a business appearing in Claude?
Not by itself. Anthropic's documentation describes ClaudeBot as collecting content for model training, with Claude-User and Claude-SearchBot handling retrieval and search quality. Blocking ClaudeBot alone opts out of training while leaving retrieval open, which can be a deliberate choice rather than a mistake.
Why does a PageSpeed score of 57 not mean the site is broken?
Because that score is one throttled lab simulation. Two of the five sites we measured had real-user field data, and their field LCP was 1.7 and 1.2 seconds, inside Google's 2.5-second threshold. Check the field numbers before authorising any rebuild.
Can I cite these figures?
Yes. Please cite "SingRank, Singapore Corporate Gift Website Study 2026" with a link to this page and the collection date, 21 October 2026. Quote each count with its base, such as 3 of 8, and keep the metasearch caveat. Challenges to a number are welcome through our contact page.