A US client is looking at your one-pager. There are six client logos across the top, three testimonials in italics, a case study with a chart that goes up and to the right, and a line that says "97% on-time delivery." Every one of those is a claim you made about yourself. None of them is checkable from where the client is sitting, ten and a half time zones away, with no shared court either of you would casually use and no cheap way to call the reference who is themselves a curated pick.

This is the portfolio problem, and it is worse in the corridor than anywhere else.

The portfolio is a highlight reel, not a record

Logos, screenshots, testimonials, and star counts share a family resemblance: they are assertions the shop makes about the shop. They are cheap to fabricate and cheaper to curate. A logo tells you a company once paid an invoice, not whether the engagement hit spec, hit dates, or ended without a dispute. A testimonial is a sentence someone agreed to have attributed to them, usually after the good projects and never after the bad ones. That selection bias is the whole point of a portfolio. You show your best three builds. You do not show the engagement where scope tripled, the client withheld the final payment, and both sides stopped replying to email.

So the portfolio answers "does this shop look impressive?" It does not answer the three questions a buyer actually has: did they hit the agreed spec, did they hit the agreed dates, and did they get through it without a fight. Those are questions about commitments and outcomes. A highlight reel is structurally incapable of answering them, because it was assembled by the party being evaluated, with full freedom to include and omit.

Now add distance. A US client evaluating an Indian shop cannot cheaply verify any of it. There is no small-claims path either party would realistically invoke across the border. References are curated and often coached. Due diligence runs into time-zone friction and two different legal systems. So conventional portfolio signals carry less weight precisely in the corridor where trust is hardest to build. The signal is weakest exactly where it is needed most. That is a bad trade for a good shop, because a good shop is the one whose real record would clear it, if only the record were legible.

What a delivery record is instead

A delivery record is a different kind of artifact. It is not a curated selection of outcomes; it is the running ledger of what was promised and what was delivered against each promise. Four components:

  • The acceptance criteria agreed up front for each deliverable. The definition of "done" that both sides signed before work started, written as concrete, checkable conditions rather than a vibe. Mature engineering already does this. Acceptance criteria are, per Cucumber's own definition, "conditions that a product must satisfy to be accepted," and teams routinely write them as Given/When/Then statements that are close to executable. The record layer does not invent acceptance criteria. It takes the ones good teams already write and makes the agreement itself durable. (More on getting these right in acceptance criteria a client can't argue with.)
  • Evidence attached to each milestone showing those criteria were actually met. Not "we finished sprint 4," but the artifact that demonstrates the agreed condition held.
  • Change orders, capturing every renegotiation of scope, date, or price at the moment it happened, not reconstructed later from memory. This is also where the record of what was actually agreed lives, which is the same thing that keeps a deal from stalling when the client's counsel starts asking what "done" means. That fight is its own subject; see why the client's lawyer is slowing your close.
  • Sign-offs from the client accepting each milestone. Their acceptance, not your assertion of their acceptance.

The load-bearing property is that this record is append-only and tamper-evident. Entries can be added. They cannot be silently edited or deleted after the fact, and any attempt to rewrite an earlier entry is detectable. That single property is what converts a claim into proof. A claim can always be revised to flatter whoever is telling it. A tamper-evident record cannot, because altering history leaves a mark. A testimonial says "they were great." A delivery record says "here is what was agreed on the 14th, here is the evidence the client accepted on the 28th, and none of it has been quietly rewritten since." One is a sentence. The other is costly to fake.

Why it is a moat for the shop, not just the client

The client benefit is obvious: they can inspect instead of trust. The part that gets missed is that the record is a moat for your shop, and for three specific reasons.

It is portable. It is proof you own and carry from one client to the next. A five-star average on somebody's marketplace is locked to that marketplace. Leave the platform and the rating evaporates; it was never yours. A delivery record travels with you. It is your credential, not the platform's asset.

It compounds. Every completed engagement adds another verified entry. The evidentiary weight grows monotonically with the work you have actually done, and because the record is tamper-evident, it cannot be bought, backdated, or backfilled. A new shop cannot conjure four years of accepted milestones. You can only accumulate them the slow way, which is exactly what makes them worth something.

It lets a good shop out-signal a cheaper one on something other than price. When all anyone can show is claims, competition collapses toward whoever claims the most and charges the least, because the buyer has no way to tell a real track record from a well-written one. A verifiable record is a credential the low-quality shop cannot cheaply counterfeit. It moves the contest from "who claims more" to "who can prove more." If you are the shop that actually ships to spec, that is the only ground you want to fight on.

A very old idea, hedged

None of this is new in shape. Private trade has repeatedly built its own record-and-reputation machinery where formal state courts could not reach or were too slow to matter.

The usual touchstones are the medieval "law merchant," or lex mercatoria, and the reputation networks among long-distance traders. In one influential model, Milgrom, North and Weingast (1990) treat the law merchant less as a body of rules and more as an information institution: a way to make a trader's past conduct legible to the next counterparty, so reputation could do work that courts could not. In a related vein, Greif (1989) describes an eleventh-century coalition of Maghribi traders that deterred agents from cheating across long distances by the shared threat of lost future business.

Both stories are genuinely contested, and it matters that they be told with the hedge attached. Emily Kadens argues in "The Myth of the Customary Law Merchant" that the idea of a uniform, merchant-made medieval law is largely a later foundation myth rather than accurate history. Greif's reputation-coalition reading is disputed too: Edwards and Ogilvie contend the Maghribis leaned on formal legal enforcement more than the informal-reputation account allows, and Greif rebuts them, a back-and-forth collected on his own page. Historians will keep arguing the specifics. The durable point, the only one worth borrowing, is narrow: cross-border trade has long needed its own trust infrastructure, and merchants have repeatedly built one when the state's did not stretch far enough.

The corridor is heading toward the record

The entire US–India problem is verification across distance and across two institutional systems. Everything else is downstream of it. A tamper-evident delivery record collapses that problem rather than working around it. Instead of the client trusting your pitch or running expensive cross-border diligence, the proof travels with the deliverable and checks itself. The buyer inspects the record; they do not have to take your word.

That is the direction the corridor is moving: from "trust the pitch" to "inspect the record." Which is also the point of view behind building a record layer at all. If proof beats claims, then the thing worth building for this corridor is the infrastructure that produces portable, tamper-evident proof of what was agreed and what shipped. Not a nicer portfolio. A different category of artifact.

Your best work is already real. The question has only ever been whether the person on the other side of the ocean can see that it is real without having to take your word for it. Build the record, and they can.