Build, Buy or Integrate: How to Choose the Right Digital Solution
Buy SaaS for solved problems, integrate when your tools just need to talk, build only where software is your edge. A practical framework with 3-year cost math and a worked example.

Obsidian Super Admin
6 min readObsidian Wolves team
It is common to find a company paying for nine SaaS subscriptions while its order process still runs on a shared spreadsheet. It is just as common to find one that spent eight months building a custom CRM that does less than an off-the-shelf product charging a few dollars per seat. Both made the same mistake from opposite directions: they picked a delivery model before they understood the problem.
Build, buy or integrate is not a philosophy. It is a decision you make per capability, and most healthy companies end up using all three. This article gives you a framework to make that call in an afternoon, with numbers you can defend to a board.
The three options, defined plainly
Buy means adopting an off-the-shelf product, usually SaaS, and changing your process to fit it. Accounting, payroll, email, helpdesk and most HR workflows belong here. These are solved problems, and the vendor spreads the cost of solving them across thousands of customers.
Integrate means keeping the tools you have and connecting them, or extending them with configuration, plugins and small amounts of glue code. Your CRM pushes won deals to your ERP, your store syncs stock with your warehouse system, your forms land in the right queue automatically.
Build means custom software that you own: your logic, your data model, your interface. It is the most expensive option up front and the only one that can become a real competitive advantage.
Six questions that decide it
Score each capability you are considering against these six questions. You do not need a spreadsheet model at first. Honest answers usually point clearly in one direction.
Differentiation. Does doing this better than competitors win you customers? If a customer would never notice the difference, buy. If the workflow is the product, such as a pricing engine, a booking experience or a logistics dispatch rule set, it is a build candidate.
Total cost of ownership over three years. Compare full costs, not sticker prices. SaaS costs grow with seats and tiers. Custom software costs concentrate in year one, then settle into maintenance.
Data ownership. Who controls the data, in what format, and how hard is it to export? If your customer data or operational history is a strategic asset, you need clean, complete access to it regardless of which option you pick.
Compliance and data residency. In Saudi Arabia, the Personal Data Protection Law and sector rules from regulators such as SAMA can constrain where personal and financial data is stored and processed. The UAE has its own federal data protection law plus free-zone regimes. Check where a vendor hosts data before you sign, not after.
Speed to value. How soon do you need this working? A SaaS tool can be live in days. A focused custom build takes weeks to a few months.
Vendor lock-in. If the vendor raises prices by 40%, gets acquired or drops a feature, what does leaving cost you? Lock-in is acceptable for commodity functions and dangerous for core ones.
Doing the three-year cost math honestly
Most build-versus-buy decisions go wrong on cost because people compare a custom build quote against one month of SaaS pricing. Put both on a three-year horizon and include the costs nobody quotes you.
For buy, include:
Subscription fees at your projected headcount in year three, not today
Tier upgrades you will need once you hit usage limits
Implementation, data migration and training
The workarounds staff build in spreadsheets when the tool does not fit
For build, include:
The initial build, including design, testing and launch
Hosting and infrastructure
Ongoing maintenance and improvements, commonly budgeted at 15 to 20% of the build cost per year
Internal time for product decisions and acceptance testing
For integrate, include the subscriptions you already pay, the integration work itself, and the monitoring needed so a broken connection gets noticed in hours, not weeks.
A worked example
Take a hypothetical Riyadh-based distributor with 40 field sales reps who take orders from shops, check stock and apply customer-specific pricing. Today it runs on WhatsApp messages and a spreadsheet the office team re-keys into the ERP.
Option A: buy a field sales SaaS. Say it costs 150 SAR per user per month. For 40 reps growing to 60 by year three, that is roughly 72,000 SAR in year one and about 108,000 SAR in year three, so around 270,000 SAR over three years, plus 40,000 SAR of setup. The catch: it cannot model the distributor's tiered pricing per shop, so the office keeps a spreadsheet for exceptions. The re-keying problem shrinks but does not disappear.
Option B: integrate. Keep the ERP, add a lightweight order form that reads stock and price lists from it. Say 60,000 SAR of integration work plus 15,000 SAR a year of upkeep, about 105,000 SAR over three years. Re-keying ends. But the pricing logic still lives in the ERP's rigid rules, and reps get a basic form, not a tool that helps them sell.
Option C: build a focused field app. Say 220,000 SAR to build an app with offline ordering, the full pricing engine and a sync to the ERP, plus 40,000 SAR a year of maintenance and hosting. About 340,000 SAR over three years. It costs the most, but it encodes the company's pricing advantage and the company owns the code.
The right answer depends on one question: is pricing flexibility how this distributor wins accounts? If yes, Option C earns its premium. If the company wins on delivery speed instead, Option B delivers most of the value at a third of the cost. Notice that the decision came from differentiation, not from the price tags.
Common mistakes
Building commodity software. Custom HR systems, custom chat tools and custom CMSs rarely pay back. Your team will spend years maintaining what a vendor already perfected.
Buying your core. If your differentiating workflow lives inside a vendor's configuration screen, your competitor can buy the same advantage next month.
Ignoring integration until launch. A new tool that does not connect to your existing systems becomes another island of data, and staff go back to copying between screens.
Comparing year-one costs only. SaaS looks cheap at 10 seats and expensive at 200. Custom software looks expensive in month one and cheap in year four.
Skipping the exit plan. Before signing any contract, confirm you can export all your data in a usable format. Before commissioning any build, confirm you will own the code, the IP and the infrastructure accounts.
Treating it as a one-time decision. A capability you bought at 20 employees may deserve a build at 200. Review the big ones yearly.
How to run the decision in practice
Start by listing your capabilities, not your tools: take orders, price quotes, onboard staff, report to management. Mark each one as parity or advantage. Parity capabilities go to buy or integrate. Advantage capabilities get a closer look.
For anything heading toward a build, define the smallest version that proves the value. A focused first release in weeks tells you more than a six-month specification. If you are weighing a build seriously, an outside view of architecture, cost and risk is cheap insurance; that is the kind of question our technical consulting work is designed for.
Then hold each option to the same standard: clear ownership of data, a realistic three-year cost, and a named person accountable for it after launch.
The takeaway
Buy what everyone needs, integrate what needs to flow, and build only what makes you different. Run the numbers over three years, check where your data lives, and make sure you can always leave. When you do decide to build, insist on owning every line of it.
If you have a capability on the edge of that decision, our software development team can help you scope the smallest build that proves it is worth doing.
