Market intelligence
Today's tender list is a commodity. The history behind it is not.
Any portal can show you what was published this morning. Almost none can tell you that this buyer runs the same tender every March, that one firm has won it four times, and that the current contract ends in eleven weeks. That is what we built.
What is in it today
These are live figures from the catalogue, and they grow every day the system runs.
Central, 36 state portals, and global sources
Each one a single record, however many ways its name is spelled
Every tender, award and notice, dated and attached to a buyer
India, EU, US, World Bank, ADB, UNDP and more
The problem with every tender list
A tender notice tells you what one buyer wants this week. It tells you nothing about whether they are a good customer, what they normally pay, how crowded the field is, or when they will buy again.
What you can see on a portal
- A notice, published today, closing in twelve days.
- The buyer's name, spelled however that department spells it.
- A document bundle you have to read to find the eligibility rules.
- Nothing about last year, and nothing about next quarter.
What the same notice looks like here
- The buyer, as one record, with every notice they have ever published.
- What they usually buy, in what value band, in which states, in which months.
- Who has been winning their work, and how often.
- Which of their contracts end soon — and therefore what they must tender next.
How it is built
Four pieces, in order. Each one is unglamorous, and each one is why the numbers above can be trusted.
One record per organisation
A portal writes “NTPC Ltd”, “N.T.P.C. Limited” and “National Thermal Power Corporation” and means one company. We resolve them to a single record — and keep the 28 state Public Works Departments apart, because those really are 28 different buyers. Without that, every number you could calculate is wrong.
Every notice becomes a dated fact
A tender, a corrigendum, an expression of interest, an award, a cancellation — each becomes one dated signal attached to a buyer. A buyer profile is then arithmetic over that history rather than a guess, and it gets deeper every single day the system runs.
Awards are read, not just listed
An award notice names who won, who lost, the signed price and how long the contract runs. Most systems file it as “closed”. We read it — which is where competitor intelligence and contract-expiry predictions come from.
It compounds, and it cannot be bought
The notices themselves are public; anyone can fetch today's. What cannot be fetched is five years of resolved history. Every day this runs the picture deepens, and a competitor starting tomorrow starts from nothing.
What you do with it
Stop guessing which buyers are worth your time
Search for what you sell and see the organisations that actually buy it, ranked by how much and how recently. Open one and you get its purchase history, typical contract size, the states it works in and the seasons it buys in.
Know who you are bidding against
Award notices name the winning bidder and the rejected ones. Over a buyer's history that becomes a share: who wins here, and how often. You learn whether a buyer is genuinely contestable before you spend a week on the submission.
Arrive before the tender does
A contract ending in March means a tender in January. An expression of interest means a tender in three to nine months, and only firms that answer it get invited. A new plant announcement means procurement for a year. All three are visible in the public record long before a tender is published — if someone is reading them.
Work the whole pipeline, not just today's list
Live tenders, pre-tender notices, contracts running out, and private projects announced in the news — ranked against your Company DNA, so you see your ten rather than everyone's two hundred.
A worked example
Nothing here is hypothetical — this is the shape of a record the system builds from published notices.
An award notice says
Contract awarded. Scope: construction of additional classrooms. Signed contract price ₹4.09 crore. Duration of contract: 1.5 years. Date of award: 21 January 2026. Awarded bidder: one firm. Rejected bidders: two more, named.
So the system knows
- This buyer pays around ₹4 crore for work of this kind.
- That firm is a competitor here, and two others bid and lost.
- Around July 2027 this contract ends, so the re-tender comes before it.
Multiply that by every award notice published, and you have a map of who buys what, from whom, and when they will need it again.
How we get the data, and what we will not do
An intelligence product is only worth having if it is built lawfully. These are the rules we hold ourselves to, and they are not negotiable.
Facts, with the source attached
We store the facts — who, what, where, how much, when — and link out to the original notice. Every record names where it came from.
Read openly or not at all
Official APIs, open data and published feeds. We identify our reader, honour robots.txt and never log in, solve a CAPTCHA or rotate identities to get at a page.
Organisations, never people
Profiles are built for companies and departments. A sole proprietor who bid under their own name is left out of the intelligence layer entirely.
A person approves every merge
Deciding two names are one organisation is where this goes wrong. The system proposes; a human approves. Nothing merges on a similarity score alone.
Facts such as a reference number, a date, a value or a buyer's name are not anyone's property, and we reproduce only those, with a link back to the notice they came from. Where a portal asks that its pages not be read automatically, we do not read them.
Upload your next tender and see what the system already knows.
Free to start. Nothing to pay. Your documents stay private to your company.