Make the internal case
The hard part of this decision is not choosing a vendor. It is getting your own organisation to agree. These are the artefacts for that — written to be forwarded, not to be admired.
A decision like this typically involves between five and sixteen people across several functions, and the research on B2B buying groups is consistent that most of them never speak to the supplier. Conflict inside the group, not the comparison between vendors, is what usually kills the decision — and it peaks about a third of the way through, which is roughly now.
So everything below is written for the group rather than for you, contains no sales language, and can be copied out of this page into your own documents. If something here is wrong for your situation, change it — it is more useful to you edited than quoted.
1. The arithmetic
What the open role is already costing
The comparison most groups make is pilot cost against doing nothing, and doing nothing is scored at zero. It is not zero. The cost of an unfilled role is whatever it is blocking, every month, and it has usually been running for longer than anyone has written down.
Fill in your own numbers. The only figure we have supplied is our rate, and it is the published band rather than a quote — $28–38/hour for a senior engineer, which is $4,480–6,080/month full time.
If B is smaller than E, offshore is not the answer to this problem and you should not buy it. That happens, and it is worth finding out on a spreadsheet rather than in month four.
Cost-of-inaction model
COST OF LEAVING THE ROLE OPEN — fill in your own numbers A Fully loaded cost of the role if filled locally, per month $ ______ B Revenue or delivery value the role unblocks, per month $ ______ C Months the role has already been open ______ D Expected further months to fill locally ______ Cost already absorbed = B x C $ ______ Cost still to come = B x D $ ______ Total cost of continuing to wait = B x (C + D) $ ______ Offshore alternative E Our band for this seniority, monthly $4,480–6,080/month F Time to someone working 2 weeks G Cost of a 3-month pilot = E x 3 $ ______ Decision line Total cost of waiting (B x (C + D)) versus pilot cost (E x 3) If B exceeds E, every month of delay costs more than the pilot.
2. The risk register
Risks, mitigations, and what is left over
These are the risks a systematic review of the offshore outsourcing literature identifies as the critical ones, not a list we chose because we answer them well. Each row says what remains after the mitigation, because a register with no residual risk is not credible in front of a security or finance reviewer — and it says how to verify the claim without taking our word for it.
| Risk | How we handle it | Residual risk | How to verify |
|---|---|---|---|
| Language and cultureMisunderstood requirements, and a team that agrees with everything rather than pushing back. | One interview stage is a working disagreement, conducted in your hours. Engineers who cannot challenge a proposal politely do not pass. Written-first working: decisions land in your tracker, not in a call. | Real. Deference is a tendency, not a defect we can certify away. Expect to ask for disagreement explicitly for the first month. | Ask us to arrange a 20-minute technical disagreement with a candidate before you sign anything. |
| Project and delivery managementWork happens, but against the wrong priorities, and nobody notices for a fortnight. | Augmented engineers take direction from your lead and sit in your standup. Pods carry our own delivery lead who is accountable for output and reports weekly in writing. | If you have no technical owner on your side, augmentation will fail regardless of vendor. In that case buy a pod or a fixed-scope project instead. | Name your technical owner before the engagement starts. If you cannot, that is the finding. |
| Intellectual propertyOwnership disputed, or unenforceable across borders. | IP assigned to you at the moment of creation — comprehensively, exclusively, irrevocably, perpetually — not on final payment. Confidentiality and assignment in every employment contract on our side. Contract governed by your law where you require it. | Cross-border enforcement is slower than domestic enforcement whatever the contract says. Mitigated, not eliminated. | Send the clause to your counsel before the pilot, not at renewal. |
| Information securityAccess that outlives the engagement; production data on an unmanaged laptop. | ISO 27001 and SOC 2 Type II. Named accounts under your identity provider, least privilege, access revoked on the last working day. Company-managed encrypted devices. No production data on developer machines. Named sub-processors and a written DPA with transfer terms for your jurisdiction. | Some workloads cannot leave your perimeter — classified work, certain regulated health and financial data, anything requiring citizenship-gated clearance. We scope those out rather than work around them. | Request the SOC 2 report under NDA and check the ISO certificate against the issuing registry. |
| Technical capabilityThe CV is senior and the engineer is not. | Roughly one engineer in twelve is accepted. You interview every shortlist before anyone starts, and the rate band tells you which seniority you are being quoted. Wrong in the first two weeks, we replace at our cost. | Interviews are imperfect. The two-week replacement term exists because of that, not instead of it. | Give the candidate a problem from your actual backlog during the interview. |
| Continuity and churnThe engineer who learned your system leaves. | Engineer retention 86%. Replacement with an overlapping handover, and documentation obligations written into the engagement rather than promised. | Churn is never zero. The question is whether the handover is contracted or improvised. | Ask what the handover looks like, in days, and get it in the agreement. |
| Hidden costThe hourly rate is not the price. | No setup fee, no recruitment fee, no per-engineer onboarding charge, no management percentage. The published band is the price; the monthly equivalent and the hours basis are both on the rates page. | Your own management time is a real cost and we do not charge for it. Budget it. | Ask every vendor for the all-in monthly figure for one engineer for twelve months. |
3. The reference call
Questions worth asking our references
A reference call where the vendor picked the reference and the buyer asks whether they were happy produces no information. These are the questions that would surface a problem if there were one, and we would rather you asked them than asked us.
Ask for a reference from an engagement that has ended. How a vendor handles an exit tells you more than how it handles a honeymoon.
Reference-call script
REFERENCE CALL — questions that would expose a weak vendor Ask the reference, not the vendor. 1. What did they get wrong in the first month, and how did you find out? 2. Who did you escalate to, and how long did it take to reach a human with authority? 3. Did you ever replace an engineer? What did the handover actually look like, in days? 4. How many hours of real overlap did you get, as opposed to what was promised? 5. Did the people who turned up match the CVs you interviewed? 6. What is on your invoice that you did not expect? 7. How much of your own team's time does this consume each week? 8. Did they ever tell you not to buy something? 9. If you were starting again, what would you put in the contract that is not in it now? 10. Would you use them for something you could not afford to go wrong? Ask us for a reference whose engagement ended, not only ones that are running. The ones that ended are where you learn how an exit is handled.
4. The memo
A draft you can rewrite
The decision gets made in a document that circulates without us. This is a starting draft — the bracketed parts are yours, and the section headed “what I am not proposing” is the one that usually does the most work in a sceptical room.
Pair it with the comparison against hiring locally and the security evidence pack, which is what the security reviewer will ask for first.
Internal memo, draft
Subject: Proposal — offshore engineering capacity for [team], 3-month pilot Context [Role] has been open [N] months. Local search has produced [N] candidates at [salary], none accepted. The backlog affected is [what], which is blocking [outcome]. What I am proposing A three-month pilot with two engineers from Skydream Digital (Mohali, India), at [rate band], roughly [monthly figure] per engineer per month, all-inclusive. 30 days' notice to stop. We interview the shortlist before anyone starts. Why this vendor rather than the alternatives - Published rate bands by seniority; no setup, recruitment or management fees. - ISO 27001 and SOC 2 Type II certified. SOC 2 report available under NDA. - Contracted overlap window with our working day, not best efforts. - IP assigned at creation, not on payment. - Replacement at their cost if an engineer is wrong in the first two weeks. What I am NOT proposing Not a replacement for local hiring, and not a transfer of ownership. Our team keeps the roadmap, the code review standard and the decisions. Risks and how they are handled See the attached risk register. The residual risks I would flag to this group are [pick two from the register], and the pilot is structured so we find out about both inside 90 days for [pilot cost]. What I need from this group Approval of [pilot cost] and a decision by [date]. If the pilot does not work, we stop at 30 days' notice and have spent [figure] to learn it.
Questions
Getting a decision made
- Why would a vendor publish the questions that could lose it the deal?
- Because the decision is not made in a conversation with us. Research on B2B buying groups consistently finds that suppliers get a small single-digit share of the buyer's time, and that most of the deliberation happens internally without us in the room. Whatever we would say on a call has to survive in a document we are not present to defend — so it is more useful to write the document properly than to protect it.
- What should this group actually decide?
- Whether a three-month pilot at a known cost is worth the information it produces. That is a smaller and more reversible decision than 'should we go offshore', and it is the one the arithmetic on this page is built around.
- Who normally has to agree?
- In our experience the engineering owner, finance, and whoever owns security — plus procurement above a threshold. The security review is the one that most often adds weeks, which is why the whole evidence pack is on the security page rather than behind a conversation.
- What if the group disagrees?
- Then the pilot is the right instrument, because it converts an argument about opinions into a bounded experiment with a cost you have already approved. Most of the disagreements we see are about risk tolerance rather than facts, and 90 days of evidence settles those better than another vendor call.
Ask for what you need to finish the case
A completed security questionnaire, a reference from an ended engagement, a rate in writing for your specific role, or a redlined MSA for your counsel. Tell us which and we will send it — no call required first.