A practical, opinionated guide for Pakistani business owners evaluating software houses in 2026. 12 questions to ask, 8 red flags to walk away from, pricing benchmarks, and how to evaluate proposals. Written by people who run a software house — and who lose deals to better-prepared buyers.
Published 20 May 2026 · By M. Ahmed Siddique · 12 min read
This is the article we wish every Pakistani business owner read before contacting their first software house. We sell software services ourselves — so yes, we are biased toward the kind of buyer behaviour that makes us look good. But we also lose to better software houses regularly, and we lose to weaker software houses regularly, and watching how Pakistani buyers select vendors has taught us patterns worth sharing.
Use this guide whether you are talking to CodeForge or to any other Pakistani software house. The patterns hold across vendors. If a vendor cannot survive these questions cleanly, walk away — including us, if we cannot survive them in your specific case.
Not screenshots. Not "we built this for a confidential client." Three real URLs you can click through to. If they cannot produce three live projects, the projects do not exist or are tiny. The vendor is either lying, or new (in which case they should disclose that), or building only behind NDAs (in which case ask why and probe carefully).
Some Pakistani software houses have brilliant salespeople and weak engineers. Some have brilliant engineers but the engineers never talk to clients. Ask to see the GitHub profile, portfolio, or LinkedIn of the actual engineer who will write your project. If they refuse, that is a red flag.
Time-and-materials engagements have their place (mostly very large, long-term builds) but for typical Pakistani SME projects, fixed-scope fixed-price is the contract structure that protects you. If a vendor refuses to commit to fixed pricing for a clearly defined scope, walk away.
Get specifics in writing. Is there a per-week penalty? Is there a refund on partial completion? Is the milestone-based payment structure aligned with deliverables (so they cannot collect 80% of payment for 30% of work done)? Most Pakistani software contracts protect the vendor; insist on terms that also protect you.
The answer must be: you do, from day one, with no licensing restrictions. Get this in writing. Some Pakistani software houses retain "ownership" of their custom code and license it to you — this is unacceptable. You paid for it. You own it.
How many days/months of free bug fixes? What is the hourly or monthly rate after that? Will the same engineers who built the project be available for support, or will it be handed to a junior team? Without clear post-launch terms, you risk being held hostage when production issues arise.
Vendors who refuse this are hiding something. Vendors who agree this is a basic step of due diligence. Use those calls to ask the real questions: did the project ship on time? Were there change-order disputes? Are bugs being fixed promptly? Would you hire them again?
Ask to see actual artefacts from a recent project — Slack channels, Trello/Jira boards, weekly status reports. Vendors with disciplined processes will show them off proudly. Vendors who improvise their process will dodge or produce something obviously fabricated.
"We test thoroughly" is a non-answer. Specifics matter: automated unit tests, integration tests, manual QA pass on which devices/browsers, performance testing, security testing if relevant. The right level of testing depends on project complexity, but the answer should not be "we just check it works".
If they cannot give you a real address, or the address is a virtual office, or it is residential — that is significant signal. Not necessarily disqualifying (some excellent solo engineers work from home), but it tells you about scale and stability. The vendor should be transparent about it.
If they cannot give you specifics ("very fast", "depends on scope") without first understanding your scope in detail, that is a red flag. Conversely, if they promise a timeline that seems too fast (1-month custom ERP, 2-week mobile app), they are either lying or planning to deliver something inadequate.
Get specifics in writing: hours per month, response time SLA, security patches included or extra, hosting management included or extra, feature additions handled at what rate. Vendors who quote "PKR 25,000/month maintenance" without specifying what is included are setting you up for surprise bills later.
For anything beyond a basic 5-page website, this pricing is impossibly cheap. The vendor is either using templates (which is fine if disclosed), going to disappear after partial work, or running a scam. Real custom development cannot be priced this low in Pakistan in 2026.
Maintenance has ongoing cost. Vendors who promise lifetime free maintenance are either lying or running a business model that cannot sustain itself. Both outcomes hurt you when you need maintenance and the vendor has vanished.
"We'll figure it out as we go" sounds collaborative but produces disasters. Get scope in writing. Vendors who refuse are setting up for scope-creep arguments later.
The domain must be registered in your name. Hosting credentials must be yours from day one. Vendors who hold these as "convenience" are positioning themselves to charge premium rates when you eventually want to switch vendors. Walk away.
Real client feedback includes some friction — projects that ran a bit long, features that needed iteration, lessons learned. Vendors whose testimonials read like marketing copy with no specifics are using fake testimonials.
"This pricing is only available this week" is a sales tactic, not market reality. Real software pricing does not change weekly. Pressure to sign quickly is designed to prevent you from doing proper due diligence.
Even if you are non-technical, you can ask basic questions about hosting, database choice, security approach, and backup strategy. The vendor should give clear answers, not deflect. If your project is non-trivial, consider bringing a technical advisor (paid or favour-based) to the proposal review.
The company should have identifiable founders or directors, ideally with public LinkedIn profiles. Shell companies with no identifiable humans are vehicles for disappearing when projects go bad. Walk away.
For sanity-checking quotes, here are realistic 2026 PKR pricing ranges:
Quotes significantly above these ranges deserve scrutiny (what extra are you getting?). Quotes significantly below deserve more scrutiny (what is being cut?).
When you have proposals from 2-3 software houses, compare them across these dimensions:
The cheapest proposal often becomes the most expensive once you account for missed milestones, scope arguments, and post-launch surprises. The most expensive proposal is not automatically the best either. Look for the proposal where you understand exactly what you are getting and from whom.
Trust your gut on character and communication. Software projects are long engagements where things will go wrong — bugs, delays, scope changes, mutual misunderstandings. The vendor's character under stress matters more than their initial sales polish. If a vendor is difficult to communicate with before they have your money, imagine what they will be like after.
Pick people you can work with for 6-12 months without the relationship deteriorating. The Pakistani software industry is too small for hostile vendor relationships — your project either succeeds with collaborative communication, or fails amid lawyer letters. Choose accordingly.