9 Salesforce Contractors, 3 Broken Orgs — the 7-Step Vetting Method I Wish I’d Had Before Our First Dubai Hire

Salesforce is the most commonly mis-hired role in Dubai enterprise technology, and the reason is structural: one product name covers three distinct professions. An administrator, a developer and a technical architect share a platform and almost nothing else — different daily work, different failure modes, and salary bands that differ by a factor of four. Over three orgs and nine contractors, every expensive failure I have seen was a mismatch at that first decision, not a shortage of talent in the market.

What follows is the method we now run, in order. It assumes you are hiring in the UAE, with UAE visa realities and UAE counter-offer dynamics.

Step 1: Decide which of the three profiles you actually need

Do this before anything else, and be honest about it. The test is simple: describe the first three things the hire will do in week one. If the answers are “build a report, adjust a page layout, fix a validation rule,” you need an administrator and hiring a developer will bore them into leaving within six months. If the answers involve an integration with a core banking system or a data migration, you need a developer or an architect, and an administrator will drown.

The failure I see most in Dubai is the third row misapplied: a company hires an expensive certified architect to run a 40-user org that needed a strong administrator and one part-time developer. The architect is competent, has nothing to architect, and resigns in eight months. The company concludes “Salesforce talent in the UAE is unreliable.” It was not.

Step 2: Audit your org before you write the job description

You cannot write an honest job description for an org you have not measured. Spend two hours pulling five numbers:

1. Apex class and trigger count, and specifically whether any object has more than one trigger. Multiple triggers on the same object is the single clearest marker of accumulated technical debt.
2. Overall test coverage, and how much of it is real assertions versus code that runs without checking anything. Coverage at 76 percent with no assertions is worse than 60 percent with good ones, because it creates false confidence.
3. Number of integrations and whether they are point-to-point or through middleware.
4. Managed packages installed, and which ones nobody remembers buying.
5. Deployment method today — change sets, unlocked packages, or a proper CI pipeline. This one predicts more about the hire you need than any other.

Then put the honest version of these numbers in the job description. Strong candidates find a messy org with a mandate to fix it genuinely attractive. Weak candidates are deterred. That is exactly the filter you want, and it costs you nothing.

Step 3: Write a job description that filters instead of attracts

Delete the certification wish-list. Replace it with two or three concrete problems the hire owns in the first ninety days, each with a measurable outcome. “Consolidate three Opportunity triggers into one handler framework with assert-based tests above 85 percent” tells a competent candidate everything about the role. “Platform Developer II required” tells them nothing except that you screen on paper.

Keep one certification as a floor filter if you receive high volume — Platform Developer I is a reasonable one — but treat it as a filter, never as a ranking signal.

Step 4: Source where UAE Salesforce talent actually sits

Four channels, with different economics:

• Consulting partners in Dubai Internet City and DIFC. Deepest concentration of hands-on delivery experience in the country. Notice periods run 60 to 90 days and counter-offer risk is the highest of any channel, because partners protect billable staff aggressively.
• In-house teams in banking, insurance, retail and real estate. Better retention profile, narrower exposure — often deep in one cloud and shallow elsewhere. Usually 30-day notice.
• The regional remote pool — Egypt, Jordan, India, Pakistan. Strong technical depth at 40 to 60 percent of UAE cost. The trade-off is time-zone overlap for release windows and, if you eventually want them in Dubai, a full visa cycle.
• Candidates already on a transferable UAE residence visa. They start weeks earlier and they know it, which is why they price at a premium. Often worth paying.

Step 5: Run a technical screen built on governor limits and debugging

Forty-five minutes, three questions. This screen has done more to improve our pass-through quality than any other change.

Question 1 — Bulkification, asked as a story

“A trigger works perfectly in testing. A data load of 5,000 records fails. Walk me through what happened and how you would find it.” You are listening for governor limits, SOQL inside loops, and whether the candidate reaches for the debug log or for guesswork. Certification study produces the phrase “governor limits”; production experience produces a diagnostic sequence.

Question 2 — A real debug log

Show them an anonymised debug log with a genuine problem in it and ask what they see. This is the single highest-signal exercise in the loop. Candidates who have lived through production incidents scan it in a particular way and narrate what they are ruling out. Candidates who have not go quiet.

Question 3 — Declarative versus code judgement

“When would you refuse to write Apex and use configuration instead, and when the reverse?” There is no single right answer, and that is the point. Strong candidates give you a threshold and a reason. Weak candidates express a preference. Anyone who says “always code” will generate maintenance work your administrator cannot touch.

Step 6: Verify with a scoped sandbox exercise and one reference question

The exercise: a free developer org, three hours, a small but realistic brief — a trigger with a handler class, bulk-safe, with tests. Pay for their time if the market is competitive; in Dubai it usually is.

Review it on test quality, not feature completeness. Do the tests assert real outcomes or just execute lines? Is there a bulk test with 200 records? Is there a negative test? A candidate who ships slightly less functionality with genuinely good tests is the stronger hire every time, and the sandbox exercise is the only place you can observe that directly.

The reference question that matters: “How did they deploy to production, and tell me about a deployment that went wrong.” The answer reveals release discipline, whether they work with change sets or a real pipeline, and how they behave under pressure. It has predicted six-month outcomes better than any other reference question we ask.

Step 7: Close with a package structured for the UAE market

Four components, and candidates compare all four:

• Base salary against the bands in step 1. Below-band offers do not fail at offer stage — they fail at month seven, when the person leaves.
• Housing allowance, stated separately. In Dubai this is read as a distinct number, not folded into total compensation.
• Visa sponsorship timing, with a committed date. Vagueness here loses candidates to employers who commit.
• Certification and training budget. Cheap for you, disproportionately valued by Salesforce professionals whose market rate is credential-linked. An architect track commitment closes candidates that money alone does not.

Counter-offer risk is highest for candidates leaving consulting partners. Address it directly before the offer, not after: ask what would have to be true for them to stay where they are. If they cannot answer, they are ready. If they answer fluently, you are the leverage in someone else’s salary negotiation.

In-house or partner? The nine-month line

Use a consulting partner for a bounded implementation with a defined end date. Hire in-house for anything you will still be changing in two years. The economics cross at roughly nine to twelve months of continuous work: below that, partner day rates beat the fully loaded cost of a hire plus visa and recruitment fees; above it, in-house wins and you keep the org knowledge inside the company.

The expensive pattern in the UAE is a three-year partner engagement with nobody internally who understands the org. If that is where you are, hire an administrator first — not a developer — and have them shadow the partner for a quarter. If you are still deciding what to build at all, our guide to building a CRM system in the UAE covers the build-versus-configure decision that precedes this one.

Employers running the same search elsewhere in the region will recognise the pattern: our colleagues cover solutions architect hiring in Singapore and English-speaking developer hiring in Tokyo, where the profile-mismatch failure mode is identical.

Frequently Asked Questions