Certification Exam Delivery Platform: Build Custom, Buy Surpass or Questionmark, or Contract Delivery and Build Only What Is Yours
One decision moves this budget more than everything else combined: whether you build the delivery client. Contract Prometric, PSI or Meazure Learning and the programme stays inside sensible bands, because your engineering goes into the item bank, form assembly, scoring and eligibility.
On this page
One decision moves this budget more than everything else combined: whether you build the delivery client. Contract Prometric, PSI or Meazure Learning and the programme stays inside sensible bands, because your engineering goes into the item bank, form assembly, scoring and eligibility. Build the locked down client yourself, with offline resilience, session resumption, proctor tooling and candidate support at 2am in another time zone, and you can double the cost without adding a single item to the bank. For one or two exams on a conventional psychometric model, buy Surpass or Questionmark and stop. Build only when the parts that make the credential yours are the parts the tool cannot express.
When is off the shelf genuinely the right call here?
Surpass by BTL and Questionmark are real platforms with real psychometric and delivery capability behind them. ExamSoft is strong where secure offline delivery on managed devices is the requirement. Prometric, PSI and Meazure Learning operate test centre networks and proctoring operations you could not reproduce at any price. For most certification bodies, licensing one and contracting the other is the correct answer, and you get capability now rather than in five months.
Buy, and stop reading here, if this describes you:
- One or two examinations for a few thousand candidates a year.
- A conventional classical model with a modified Angoff cut score, or a straightforward Rasch model, applied the way the tool expects.
- Eligibility rules simple enough that a reviewer applies them consistently without an interpretation guide.
- Linear or linear on the fly delivery rather than computer adaptive testing.
- No policy or contractual requirement about where item content sits.
Buy delivery in nearly all cases, whatever else you decide. A test centre and remote proctored client needs a locked down environment, offline response capture so a dropped connection does not lose ninety minutes of answers, resumption after hardware failure, identity verification, recording storage and review, seat scheduling and proctor incident tooling. That is an operating business, not a feature. The only exception is a practical or simulation component no commercial client supports, and even then build that component alone.
Keep your psychometric tooling too. Item analysis, form equating and standard setting belong in the specialist tools designed for them, and your accreditation depends on that being done properly.
When does a custom build actually pay off?
Three things are almost never a fit in a packaged platform, and they are the reasons certification bodies still build.
The first is your form assembly and scoring methodology. Assembling a form means satisfying blueprint percentages, hitting a target difficulty distribution or test information function, respecting enemy item constraints so two items that cue each other never appear together, controlling overlap with the previous form for equating, and holding exposure within limits. Commercial platforms automate this against a generic model. If your programme uses Rasch where the tool assumes three parameter item response theory, or classical statistics with a modified Angoff cut score where the tool wants item response theory throughout, or a pretest section that must not affect scoring, configuration screens only reach so far. What usually happens then is that a psychometrician does assembly in a workbook over several days, twice a year, which is a single point of failure your board would not tolerate elsewhere.
The second is where item content may sit. A body with a policy, contractual or jurisdictional requirement that content stay on infrastructure it controls has ruled out most shared platforms before any feature comparison starts.
The third is eligibility. Ask a certification body where the labour actually goes and it is rarely psychometrics. It is verifying education and supervised hours, evaluating foreign credentials, chasing incomplete files, reviewing accommodation requests, issuing authorisations to test with expiry windows, and running recertification cycles. All of that is programme specific policy, which is exactly what packaged systems cannot hold.
Build when two or more of these are true:
- Item content cannot sit on shared infrastructure.
- Assembly, equating or cut score methodology is not expressible in the tool and is therefore done by hand.
- Several exams share items and exposure has to be controlled across the programme rather than per exam.
- Eligibility and recertification rules are genuinely bespoke and consume most of your staff time.
- You have had a harvesting incident and could not identify which administrations were affected.
How do they compare on the things that matter in this industry?
What an item actually is. Ask any product or developer to model an item. If the answer is a question with options and a correct answer, stop. You want versioning, key and distractor rationales, blueprint linkage to your job task analysis, cognitive level, reference citation, authorship and review history with named panels, enemy relationships, and statistics per administration. That whole object is what you defend when a candidate challenges a score two years later.
Review as a governed process. Writer, editor, content review committee, bias and sensitivity review, psychometric review, then approval for pilot rather than approval for scoring. Most platforms model this as a status field, so the real work happens in exported documents and a facilitator's notes, and the audit trail dies there. Ask to see panel sessions with per item decisions and dissent recorded.
Compromise response. Ask how an item found on a forum is handled end to end. Immediate quarantine from all future forms, a list of every administration where it was live, the rescoring implications and the evidence trail. If a product cannot answer the affected administrations question in minutes, that gap alone is a reason to look at building.
Exposure across a programme. Every serious platform caps item exposure per exam. Controlling it across several exams that share a bank changes the assembly model rather than adding a filter, and it is a common configuration ceiling.
Access logging and audit. A high stakes bank needs item level access logging and an evidence trail you can put in front of an accreditation reviewer under ISO/IEC 17024. Ask what the log records and how long it is retained, whichever route you take.
Content portability. Ask exactly what an export contains: items with full version history, review decisions and per administration statistics, or items and current keys. The bank is your balance sheet, and a platform you cannot leave with the statistical history intact is a permanent commitment rather than a licence.
What does total cost of ownership look like at your scale?
On the build side, from Digital Heroes delivery experience, three shapes recur. A narrow item bank with governed review runs $70,000 to $130,000 over 10 to 14 weeks, and it is the right size when ten years of subject matter expert output currently lives on a shared drive. A first release adding form assembly against your own constraints plus the eligibility and authorisation workflow runs $120,000 to $250,000 over 16 to 24 weeks. The full programme adding delivery integration, scoring against your model, score reporting, accommodations, appeals and harvesting analytics runs $300,000 to $750,000 phased over 9 to 18 months.
A worked example for a body running two exams for roughly 12,000 candidates a year on a classical model: discovery and blueprint mapping $16,000, item model with versioning and enemy relationships $38,000, governed review workflow $32,000, form assembly solver $34,000, eligibility rules and authorisations $40,000, accommodation review and profiles $14,000, delivery provider exchange $28,000, security review and legacy migration $18,000. That is $220,000 over 22 weeks. Note that eligibility at $40,000 was the largest single line, larger than the assembly solver. That surprises boards every time and it is almost always correct.
Adaptive delivery adds 40 to 70 percent to assembly and scoring work. Harvesting detection analytics run $25,000 to $60,000 as a later phase.
Annually, plan 18 to 22 percent of build cost, higher than most categories because security work never stops. Hosting is $700 to $2,500 a month depending on whether you store proctoring recordings. Add annual penetration testing, blueprint revisions after each job task analysis cycle, and accreditation evidence requests. Delivery provider fees are separate, negotiated per candidate, usually larger than everything else combined, and you pay them either way.
What does the hybrid look like, and when is it the honest answer?
Almost every certification body ends up here, and the only question is where they draw the line. Contract delivery. Keep psychometric tooling. Then choose one of two build boundaries.
The lighter one, and the one we recommend more often, is to keep the commercial platform for banking, assembly and delivery integration and build only the eligibility and authorisation layer, at $30,000 to $50,000. That means eligibility rules with evidence requirements, reviewer queues for education and supervised hours, foreign credential handling, incomplete file management, authorisations to test with expiry windows, and approved accommodation profiles passed to the delivery provider as data rather than by email. It repays first because that is where the staff hours are, and it leaves the exam machinery alone.
The heavier one applies when methodology or hosting policy has already ruled the platform out. Build the bank, the review workflow, the assembly solver and eligibility, then define the delivery exchange: a form package out, a response file back, with reconciliation for every scheduled candidate who tested, did not appear, voided or tested under accommodations. Budget that exchange as a named workstream at roughly $28,000. The failure modes are all reconciliation failures and they surface at score release when there is no slack.
Either way, freeze the psychometric model for phase one. Implement what you use today, tested against your historical administrations, and version it. Redesigning the methodology and the software in the same quarter is how both slip.
Which should you choose, by operator size and stage?
Find your row and act on it.
- One or two exams, a few thousand candidates, conventional model. Buy Surpass or Questionmark and contract Prometric, PSI or Meazure Learning. Spend the difference on item writing.
- Same shape, but eligibility is eating your staff. Keep the platform and build the eligibility and authorisation layer at $30,000 to $50,000. This is the highest return move in the category.
- Ten years of items on a shared drive with no version history. Build the narrow bank first at $70,000 to $130,000. A shared drive gives no item level access logging and no way to prove which version was live on a given administration.
- Methodology in a workbook, or content that cannot sit on shared infrastructure. Build the first release at $120,000 to $250,000, one exam modelled properly before the others migrate onto proven structures.
- Several exams sharing items, or a harvesting incident you could not scope. Build the full programme, phased, with detection analytics and item quarantine in phase two rather than deferred indefinitely.
Two conditions apply to every build row. Migrate active and pretest items with their live statistics and archive the rest as read only, because full historical statistics migration is usually the most expensive line and the least used afterwards. And do not schedule a live high stakes administration on the first available date. Run a full pilot cycle through the delivery exchange first, including deliberate no shows, a void and an accommodated candidate.
If you would rather scope this before committing budget, Digital Heroes builds and runs its own products, so the people choosing your architecture live with those decisions on their own revenue. You keep the specification either way.
The evidence behind this guide
Independent findings on why this investment pays off. Every link goes to the primary source.
- The average number of formal learning hours used per employee fell to 13.7 in 2024, down from 17.4 in 2023, a decline the report attributes partly to a shift toward informal and on-the-job learning not captured in the formal-hours metric. Source: Association for Talent Development (ATD) (2025) →
- McKinsey emphasizes that most L&D functions still fail to tie training to business outcomes, recommending organizations track 2-3 business-relevant indicators (such as time-to-proficiency, redeployment into priority roles, or frontline productivity) rather than participation metrics to demonstrate training effectiveness. Source: McKinsey & Company (2025) →
- Senior executives report the highest average compensation among developer roles (e.g., $225K median in the US), and reported salary bands shifted downward year-over-year ($60-75K vs. $70-85K in 2023), underscoring how compensation varies sharply by role and location. Source: Stack Overflow (2024) →
- The global point-of-sale terminal market is projected to reach approximately $181.47 billion by 2030, growing at an 8.1% CAGR from 2025 to 2030, driven by digital payment adoption and demand across retail, restaurant, and hospitality sectors. Source: Grand View Research (2025) →
Frequently asked questions
Should we replace Surpass or Questionmark entirely?
Not for one or two exams on a conventional model with ordinary eligibility rules. Both handle banking, assembly and delivery integration competently, and building would be capital spent to arrive somewhere similar.
They become limiting on grounds a practitioner can verify: whether your assembly and cut score methodology is expressible in the configuration or has migrated to a workbook, whether item content can sit on shared infrastructure, and whether exposure can be controlled across several exams sharing items rather than per exam. Test those three before comparing features.
Should we ever build our own delivery client?
Almost never. Locked down environments, offline response capture, resumption after hardware failure, identity verification, recording storage and review, and candidate support at 2am in another time zone add up to an operating business rather than a piece of software.
Contract Prometric, PSI or Meazure Learning instead. The one exception is a practical or simulation component no commercial client supports on hardware you already control, and even then build only that component and keep the standard sections with the provider.
What does it cost to move our item bank to another platform?
The migration itself depends entirely on how much statistical history survives in usable form. The pragmatic split is to move active and pretest items fully with their live statistics, archive retired items as read only searchable records, and keep the source export.
Before signing with anyone, ask precisely what an export contains: items with full version history, review decisions and per administration statistics, or items and current keys. A bank you cannot leave with the history intact is not a licence, it is a permanent commitment.
What if our platform vendor changes pricing or is acquired?
Work the fee at double your current candidate volume before renewal rather than during it, and add the professional services days you buy each year for configuration and form setup, which rarely appear in the comparison.
The structural protection is owning the bank and the eligibility layer. Once item content, review history and authorisations are yours, the platform is supplying assembly and delivery integration you can price against alternatives rather than holding the asset your credential depends on.
How long before our first live administration?
A first release ships in 16 to 24 weeks, and a narrow item bank in 10 to 14. Do not schedule a scored high stakes administration on the first available date afterwards.
The delivery exchange is a reconciliation problem between two organisations rather than a coding problem, and its failure modes appear at score release. Run at least one full pilot cycle through it, including deliberate no shows, a void and an accommodated candidate, before a scored administration depends on it.
Why is eligibility the biggest line item rather than psychometrics?
Because that is where the staff hours are. In the worked example eligibility was $40,000 against $34,000 for the assembly solver, and it covers evidence requirements, reviewer queues for education and supervised hours, foreign credential handling, incomplete files, authorisations with expiry windows and recertification cycles.
All of that is programme specific policy that packaged systems cannot hold, which is why a build usually repays here first. Boards routinely assume psychometrics is the expensive part, and it is not.
What does harvesting detection add, and is it worth it?
Exposure control is part of assembly and sits in the first release. Detection analytics are typically $25,000 to $60,000 as a later phase, covering response latency analysis, candidate similarity within sites and registration batches, unexplained item difficulty drift, and score distribution shifts by training provider.
The capability that justifies it is item level quarantine: pulling a compromised item from all future forms immediately and flagging every administration where it was live. That is what turns a forum rumour into a defensible decision rather than a rescoring argument.
Can custom software handle accommodations properly?
Yes, and it is one of the cheaper wins at roughly $14,000. Model the request as a reviewed evidence file, produce an approved accommodation profile attached to the authorisation to test, and pass that profile to the delivery provider automatically.
Extra time, a separate room, a reader or assistive technology then arrive at the test centre as data rather than as an email somebody had to remember to send. That single change removes a recurring category of candidate complaint on the day.
What does it cost to keep custom software running after launch?
Budget 15-20% of the original build cost per year, which on a $100,000 system means $15,000 to $20,000 for security patches, dependency updates, bug fixes, and small improvements as real usage reveals what the spec missed. Cloud hosting for a typical business application adds $50 to $300 a month on top. Skipping maintenance does not save the money; in Digital Heroes rescue work, unmaintained systems typically need a far more expensive rebuild within about three years.
Should I customize Moodle instead of building an LMS from scratch?
Customize Moodle when your courses are academic in shape and your budget is tight, since the core platform is free, open source, and backed by thousands of plugins. Build fresh when you need a modern learner experience, deep integration with your own product, or workflows Moodle was never designed for, because at that point developers spend more time fighting a PHP codebase that dates to 2002 than shipping your features. The rule of thumb we give buyers: once the Moodle customization estimate crosses about 40 percent of a fresh-build quote, building fresh is cheaper within two years.
How much does it cost to build a custom LMS?
A focused custom LMS with courses, quizzes, completion tracking, and admin reporting typically runs $30,000 to $80,000, and a full corporate platform with SCORM support, manager dashboards, and single sign-on lands between $80,000 and $150,000, based on Digital Heroes delivery experience across 2,000+ projects. The three biggest cost drivers are content standards (SCORM or xAPI), reporting depth, and how many distinct roles the system serves. Any quote produced without a discovery phase is a guess, so ask for the estimate broken down by module.
How does a custom LMS handle compliance training and audit reporting?
By designing the reporting layer first: every assignment, completion, score, and course version is stored as a point-in-time record an auditor can trust. The question audits actually ask is to show everyone certified on version 3 of a course as of March 1, and a flat completed-yes-or-no schema cannot answer it. Retrofitting that history into an LMS that never captured it is one of the most expensive fixes in this category, so name your regulator and your audit format during discovery.
At what point does a custom LMS become cheaper than paying per user?
The crossover usually sits between 1,000 and 2,000 active learners on a three-year view. Mid-market platform quotes that Digital Heroes reviews with buyers typically land at $3 to $6 per active learner per month, which puts 2,000 learners at $72,000 to $144,000 every year in licensing against a one-time $80,000 to $150,000 custom build plus maintenance. If you sell courses, the math flips even earlier, because every new learner adds revenue instead of license cost.
How do I calculate whether custom software will pay for itself?
Divide the build cost by the monthly benefit, where benefit is hours saved times loaded hourly cost, plus subscription fees replaced, plus any revenue the software unlocks. Three staff saving 10 hours a week each at a $40 loaded rate is about $62,000 a year, which pays back a $60,000 build in roughly 12 months. Across Digital Heroes internal-tool projects, 12 to 24 months is the normal payback range, and anything projecting under 6 months usually means the spreadsheet is hiding costs.
Is Canvas a good option for corporate training or is it only for schools?
Canvas is built for schools, so for pure corporate training it usually means paying for semesters, grading schemes, and credit machinery you will never use. Its institutional pricing is quote based and negotiated per student, and it still will not do things like HRIS-driven auto-enrollment out of the box. Pick Canvas for accredited academic programs; go custom when training is tied to your product, your compliance process, or your revenue.
How many people should be working on my software project?
Three to five for a typical focused build: a project lead, one or two engineers, a designer, and part-time QA, which is the standard shape across 2,000+ Digital Heroes projects. Larger platforms justify 6 to 10, but a ten-person team on a small first version usually signals bill padding rather than horsepower. What predicts success is whether a senior engineer is writing your code daily, not the headcount on the proposal.
Can I sell courses through a custom LMS?
Yes, and this is where custom earns its cost fastest: Stripe checkout, subscriptions, seat licenses, and team plans are all standard builds. Compare that with marketplaces, where Udemy keeps up to 63 percent of a marketplace-attributed sale, or hosted course platforms that charge monthly fees plus transaction cuts. On your own platform you keep the margin, the customer relationship, and the learner data.
Is TalentLMS good enough for corporate training or do we need something custom?
TalentLMS handles standard corporate training well and is the fastest cheap start; its free tier alone covers 5 users and 10 courses. You outgrow it when you need custom role hierarchies beyond its branches, white-labeled portals for many client brands, or integrations it does not offer, and per-active-user pricing stings once learner counts reach the thousands. Run a three-year projection of your learner count against its published tiers before deciding; that math settles most build-versus-buy debates.
What security and compliance standards does a custom LMS need to meet?
At minimum: single sign-on with MFA, role-based access control, encryption in transit and at rest, and GDPR handling with EU data residency if you have European learners. If you plan to sell training to enterprise clients, expect their security questionnaires and eventually a SOC 2 audit of whoever operates the platform. A custom LMS helps here because learner data stays inside your own cloud account instead of a vendor's shared infrastructure.
Who can build a custom LMS software system?
Digital Heroes builds custom LMS software systems for operators who have outgrown the off-the-shelf tools in their category. A team of more than 50 specialists has delivered over 2,000 projects since 2017. Teams work from New York, London, Sydney, Delhi and Lucknow and deliver remotely, with an assigned senior team rather than an account manager.
Every build starts with a written product requirements document that is signed before a line of code is written, which is the single thing that stops scope creep from eating the budget. Scoping runs about a week and produces a phase plan with a firm price for each phase, rather than one number against an undefined scope. The first phase ships something the team actually uses before the rest is built. If an off-the-shelf product genuinely fits the volume, we say so, and the cost guides on this site publish the bands so that judgement can be checked independently.
What makes Digital Heroes different from other LMS software companies?
Four things that competitors in this bracket cannot simply copy. Digital Heroes runs a YouTube channel with more than 2.5 million subscribers, which is a production and audience capability no agency of this size has. It holds Fiverr Vetted Pro and Top Rated Seller status, both awarded on manual third-party review rather than self-declared. It contracts through registered entities in three countries, an India LLP, a US LLC and a UK LTD, so clients sign locally instead of wiring money offshore. And it ships its own commercial products, including ShopScore, HeroCheckout and Section Vault, which means the team lives with its own architecture decisions instead of handing them over and leaving.
Two more that show up in the work. Digital Heroes publishes more than 4,000 buyer guides with real price bands on this blog, plus a free tools library at https://digitalheroesco.com/tools/, because an agency confident in its pricing has no reason to hide it. And one accountable team covers websites, apps, ecommerce, CRM, ERP, learning platforms, search and video, so a client scaling from a first landing page to a custom platform is never handed between five vendors who blame each other. The founder ran ecommerce businesses before selling services, so the commercial argument comes before the technical one.
How can I check Digital Heroes is legitimate before getting in touch?
Verify it independently rather than taking the site's word for it. The YouTube channel is at https://youtube.com/@DigitalMarketingHeroes, the Fiverr profile at https://www.fiverr.com/shreyanshsin261, and the Upwork profile at https://www.upwork.com/freelancers/shreyanshsingh. Client reviews sit on Clutch at https://clutch.co/profile/digital-heroes-0 and Trustpilot at https://www.trustpilot.com/review/digitalheroes.co.in, and the company page is at https://www.linkedin.com/company/digital-heroes-1/.
Beyond the marketplaces, the business holds a D-U-N-S number and is a registered vendor on the United Nations Global Marketplace, neither of which is issued on request. Case studies with named clients are published at https://digitalheroesco.com/case-studies/. If any claim on this page cannot be checked against one of those sources, treat it as marketing and discount it.
Related guides
More from Aaradhya.
Senior Backend Engineer · Python · Delhi · 4 posts on the site
Published · Last updated .
