Calgary buyers usually arrive with a procurement habit from heavy industry
Software purchasing in this city is shaped by an engineering and resources economy, and that inheritance is visible in how projects are run. Buyers here are often comfortable with formal statements of work, milestone payments, change orders and vendor qualification, because that is how capital projects have always been contracted. That discipline is an advantage. It also imports one habit that works badly for software: the belief that a complete specification can be written before work begins, and that deviation from it is a failure of the supplier.
Physical projects reward that assumption because concrete does not change its mind. Software rewards the opposite, since most of what you learn about a system becomes visible only once people use it. The practical consequence is that the strongest engagements in this market keep the procurement rigour, the documented scope, the named accountabilities, the change log, and pair it with a delivery model that expects the specification to evolve in the early phases.
What the commodity cycle does to your supplier
This is the local risk that outside buying guides never mention. Development firms here have client bases weighted toward energy, engineering services and the businesses that supply them. When the sector is spending, the same firms are hiring aggressively and your project competes for their senior people. When it contracts, budgets are cut quickly and suppliers lose several clients at once.
So ask a question you would not ask elsewhere: how concentrated is your client base, and what happened to your team through the last downturn. You are not looking for a particular answer, you are looking for awareness. A firm that has survived a cycle will talk about it plainly and will usually have deliberate diversification. A firm that has only grown will be surprised by its first contraction, and your maintenance contract is the thing that gets cut when they need to protect payroll.
Field, plant and office software are three different problems
A large share of Calgary requirements involve people who are not sitting at a desk. That single fact reorganises the whole evaluation. Field software has to work where connectivity is unreliable, which means genuine offline capability, conflict resolution when two people edit the same record, and a synchronisation model that survives a truck driving out of coverage mid form. Plant and operations software has to integrate with control systems, historians and equipment that speak industrial protocols rather than modern interfaces. Office software has none of those constraints and is dramatically cheaper.
Suppliers who mainly build web applications will quote field requirements as if they were web applications, because from a feature list they look identical. The difference shows up in testing and in the first month of real use. Ask any candidate to describe how their proposed system behaves with no connection for a full shift, and who resolves the conflicts afterwards.
Research credits, IP and who is entitled to claim what
Bespoke development often involves technical uncertainty, and Canadian buyers and suppliers both have an interest in scientific research and experimental development tax incentives. That creates an unusual contractual question that is worth settling in writing before the first invoice.
If the supplier is carrying out experimental work at their own risk, they may intend to claim against it. If you are paying for the work and expect to own the result, you may intend to claim too. Both cannot be true in the same way, and the answer interacts directly with your intellectual property clause and with whether the engagement is structured as contract research or as a sale of a deliverable. It is not an exotic issue in this market, it is routine, and the only expensive version of it is the one discovered by an accountant a year later. Raise it during contracting, keep the documentation of who did what, and make sure the ownership clause and the tax position tell the same story.
Statements of work that survive a change of scope
Given the local preference for formal documents, the useful investment is in making the statement of work honest rather than making it longer. Three elements carry most of the weight.
The first is acceptance. Write what a completed deliverable looks like, who signs, how long they have to review, and what happens if they do not. Silent acceptance clauses prevent projects stalling behind an absent approver. The second is change. Agree the format of a change request, who can raise one, how it is priced and whether it moves the date. Change is not a sign of failure, an undocumented change is. The third is assumption. Every estimate rests on assumptions about your availability, your data quality and access to systems you do not control. List them explicitly, because an assumption that turns out false is the legitimate basis for a change and an unlisted one becomes an argument.
Choosing a pricing model instead of inheriting one
Milestone based fixed pricing fits well defined replacement work and sits comfortably with local finance processes. It requires a real specification and it makes change expensive by design. Time and materials with a cap fits discovery, integration and anything touching equipment or third party systems, and it needs you to provide a decision maker who responds within days rather than weeks.
Dedicated team arrangements fit long running product work where the backlog outlives any single scope. Their risk is drift, so pair them with a quarterly review that asks whether the outcome still justifies the run rate. Whatever the structure, agree what happens to unused budget and to work in progress if you stop early, because energy linked businesses stop projects more often than most.
Turning a longlist of Calgary software companies into a decision
Three suppliers briefed identically will tell you more than eight briefed loosely. Ask each for their read of the scope, their assumptions, their exclusions, the named people who would do the work and one example of a project that went badly and how it ended. The last question separates firms that have delivered from firms that have presented.
Check references from the operations side rather than from the executive who signed. The person who used the system every day for a year knows whether support requests were answered. If your requirement spans customer facing channels as well as internal systems, it is usually better to engage separate specialists, whether mobile app developers for field tools or web development firms for the public layer, than to ask one supplier to be strong at everything. The national and international picture sits in the software development company directory, and if local depth runs out for a specialised requirement, supplier pools in Vancouver and Toronto are close enough in time zone to work with normally. When your scope is written, put it in front of several verified suppliers at once and compare answers to the same question.