Build vs Buy: Field Data Collection and Enumerator Management for NGOs
Buy, or rather use the free tools, until you pass roughly eighty enumerators per round. KoboToolbox and SurveyCTO solve offline form capture properly and nobody should rebuild that.
On this page
Buy, or rather use the free tools, until you pass roughly eighty enumerators per round. KoboToolbox and SurveyCTO solve offline form capture properly and nobody should rebuild that. Build the layer they do not cover, meaning enumerator assignment, back check sampling, fabrication detection and donor indicator pipelines, at $70,000 to $150,000 for a first release across twelve to eighteen weeks.
The strongest argument in this category is for buying nothing
KoboToolbox is free. Open Data Kit is free. Both work offline on cheap Android tablets, both speak XLSForm so your questionnaire is portable between them, and both are maintained by people who understand humanitarian work. That is an unusual starting position for a build versus buy decision, because the incumbent costs nothing and does its job well.
So the first honest answer for a lot of organisations is that the money belongs somewhere else. If you run two assessment rounds a year with twenty five enumerators in one or two languages, and a supervisor can genuinely review incoming submissions each evening, you do not have a software problem. You have a supervision practice that works at your scale, and a custom platform would be a maintenance obligation your next funding cycle cannot carry. Spend the budget on enumerator training and on paying supervisors properly. We tell organisations this on discovery calls more often than we quote them.
SurveyCTO earns its licence when data quality is the binding constraint and your rounds are moderate. Its text and audio audits give you a real view of how an interview progressed, which is more than a spreadsheet review will ever produce. CommCare is the right buy when the work is longitudinal case management with returning visits rather than one cross sectional round. Each of these products was designed by people who have sat in a field office, and treating them as inadequate because they are not custom is a mistake that costs a lot of money to make.
Where the free tools stop and a build becomes rational
Every one of those platforms ends at the submission. They accept a form from a device and store it. What they do not know is that Amina is an enumerator on team six who is owed payment for sixty three completed interviews at a per interview rate, that her supervisor must back check ten percent of her work within forty eight hours, that four of her submissions fall outside the sampled cluster boundary, and that her output feeds an indicator whose donor wants it disaggregated by sex, age band and disability status using the Washington Group Short Set.
That layer is where the build case lives. A large round is a temporary labour operation: you recruit, train, test, assign, supervise, replace and pay several hundred people in a few weeks, and none of that is form capture. Assignment determines coverage, coverage determines whether your sample is defensible, and today assignment lives in a spreadsheet that one person maintains.
Fabrication is the second driver. Curbstoning, meaning an enumerator completing a questionnaire without conducting the interview, produces signals you already collect: interview duration far below the realistic minimum, question level timings with no pauses, location points clustered at a tea shop rather than distributed across a settlement, implausibly low variance across a whole day. Detecting those after demobilisation is documentation, not quality control. Turning them into a back check sample that reaches a supervisor's device while the team is still in the district is a workflow, and no capture tool ships one.
The third driver is reporting. When the same underlying data feeds three or more donor templates with conflicting indicator definitions, your analyst rebuilds the pipeline every quarter. Making indicator definitions versioned configuration rather than spreadsheet formulas is exactly the kind of thing worth owning.
Costing both paths against a grant cycle
The buy side is cheap in licence terms and expensive in people. Kobo costs nothing to nothing much depending on hosting. SurveyCTO sits in the low thousands per year for most country programmes. What you actually pay is the data manager who spends the last week of every round reconstructing quality checks in pivot tables, plus the analyst week per round spent coding open responses by hand, plus the occasional cost of a district's data being unusable because nobody noticed a team was inventing rows until the team had moved on.
The build side prices predictably. A first release covering enumerator records and assignment, work package distribution to devices, offline submission with paradata capture, automated back check sampling and duplicate and outlier flagging runs $70,000 to $150,000 and ships in twelve to eighteen weeks. A full platform adding enumerator payment with supervisor approval, multi language and multi script form governance, separated personal data handling, open response coding and a versioned indicator pipeline with donor template exports runs $180,000 to $450,000 phased across seven to twelve months.
Then the constraint nobody mentions in a feature comparison: the funding mechanism, not the engineering, sets your timeline. Most institutional donors treat software development as a capped or ineligible cost inside a twelve month project grant, and a build that runs eighteen weeks plus ongoing maintenance does not fit neatly into a grant that closes in month twelve. Organisations that get this built successfully fund it from unrestricted reserves, from a multi year framework agreement, or from a consortium arrangement where several members share the asset. If you have not settled which of those applies before you scope, you will scope something you cannot pay for in month nine.
Four costs that never make it into the proposal budget
The first is right to left and non Latin script rendering. It is treated as a translation task and it is not. Numeric inputs inside a right to left form with skip logic break in ways that only appear on one specific tablet model, response option order changes meaning when translated, and enumerators enter whichever numeral form their keyboard offers. Device level rendering tests belong in the release checklist, and they take real time.
The second is offline behaviour that survives reality. A device that has been offline for five days with a battery that died in between is the normal case, not the edge case. Local store, sync queue, an explicit conflict rule and a visible sync state a field officer can check are engineering, not a configuration setting. Anyone who describes offline support as caching will lose your submissions.
The third is personal data separation. Household surveys hold names, locations, household composition, income, sometimes protection incidents. Splitting identifiers from analytical data at rest, restricting the identified layer to named roles with access logging, recording consent as a structured fact including the language used and the version of the consent text, and setting retention with automatic disposal are all work items. Skipping them is the cheapest way to create an incident involving people in vulnerable circumstances.
The fourth is maintenance. Budget fifteen to twenty percent of build cost annually. In a sector funded round to round, a platform nobody is paid to keep running is a platform that is quietly abandoned within two funding cycles, and then re-procured at full price by whoever inherits it.
A test you can run against your last round
Pull the records from your most recent assessment and answer five questions with numbers, not impressions. How many enumerators worked the round? Under thirty, do not build. Above eighty with assignment managed in a spreadsheet, the case is open.
Second, do you pay per interview, and did payment disputes cost anyone real time? A per interview rate with no system link between rejected back checks and payment is the single most common source of both disputes and fabrication incentive.
Third, have you ever found invented data after the team demobilised? If yes, ask how you found it and how long it took. That answer tells you whether detection is happening in time to matter.
Fourth, count the donor templates your data feeds and check whether any two define the same indicator differently. Three or more with conflicting definitions means an analyst is rebuilding the pipeline quarterly, forever.
Fifth, name who can currently see identified household records. If the honest answer is that the analysts can, that is a protection exposure with a person's name on it, and it is worth fixing whichever way the build decision falls.
Two or more pointing toward build is a real case. One is a process fix. None means keep Kobo, keep the supervisor, and go and run the round.
Sequencing the first six months
If you build, the sequencing rule is absolute: do not rebuild the form engine. Keep KoboToolbox or SurveyCTO for capture, consume their submissions, and build the workforce, quality and indicator layers around them. That decision alone typically removes about a third of the budget and all of the highest risk engineering, and it keeps your questionnaires portable if the organisation later changes direction.
Run the new layer alongside your existing tool for one full round in parallel before it becomes the system of record. That parallel round is where you discover the assignment and back check rules nobody has written down, and where a supervisor tells you the workflow assumes a district office that does not exist. Cutting over during an active round is how organisations lose data, and we advise against it even under funder pressure.
When selecting a partner, ask them directly whether they would replace or wrap your capture tool. Anyone who proposes a new form renderer is inexperienced or padding. Ask how devices behave after five days offline and what happens when the same submission arrives twice, and expect a specific conflict rule rather than reassurance. Ask how identifiers are separated from analytical data and who can see which layer.
Digital Heroes builds from a written product requirements document before any code exists, which in this sector means the quality rules, the back check sampling logic and the indicator definitions are agreed on paper with your data manager rather than discovered in week ten. The team is fifty plus people across more than 2,000 delivered projects, holds Fiverr Vetted Pro status, and works in the open with 2.5 million subscribers at Digital Marketing Heroes on YouTube. Contracting runs through an India LLP, a US LLC or a UK LTD so intellectual property assignment sits under your own jurisdiction, and the client owns the repository from the first commit, which matters when the system outlives the grant that paid for it.
When the shortlist is down to two and you need a tiebreaker, Digital Heroes writes a product requirements document before any code exists, so the scope is fixed and priced rather than discovered later at a day rate. The document is yours whichever way you go.
The evidence behind this guide
Independent findings on why this investment pays off. Every link goes to the primary source.
- Brands not sending push notifications can lift 90-day app retention by 190%, and forfeit roughly 95 cents of every dollar spent on user acquisition when opted-in users receive no messages within 90 days; rich notifications with images see 56% higher direct open rates. Source: Airship (2024) →
- Push notification opt-in rates vary sharply by category and platform (e.g., Business apps 56.7% Android / 46.3% iOS; Games 27.8% / 20.6%); average all-category retention was 28.29% at 1 day, 17.86% at 7 days, and 7.88% at 30 days, and apps sending onboarding messages saw 24% higher install-to-purchase conversion. Source: OneSignal (2024) →
- 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) →
- Companies in the top quartile of McKinsey's Developer Velocity Index had 2014-18 revenue growth four to five times faster than bottom-quartile peers, showing that software-building capability is a driver of business performance, not just a support function. Source: McKinsey & Company (2020) →
Frequently asked questions
How much does a custom enumerator management and data quality platform cost?
A first release covering enumerator assignment, work package distribution, offline submission with paradata, automated back check sampling and outlier flagging runs $70,000 to $150,000 across twelve to eighteen weeks. A full platform adding payment with supervisor approval, multi script form governance, separated personal data handling and a versioned indicator pipeline runs $180,000 to $450,000 over seven to twelve months. Keeping your existing capture tool removes roughly a third of that.
Should we replace KoboToolbox or build around it?
Build around it in almost every case. Kobo and Open Data Kit solve offline capture properly and the XLSForm standard keeps questionnaires portable, so rebuilding that spends budget on a solved problem. What is missing sits after the submission: assignment, payment, back check sampling, fabrication detection with a supervisory workflow, and a pipeline from raw rows to donor indicator tables. That is the layer worth owning.
How long does this take, and can it be ready for our next round?
A first release ships in twelve to eighteen weeks, which usually means the round after next rather than the imminent one. Plan to run the new layer alongside your existing capture tool for one complete round before it becomes the system of record, because that parallel round surfaces the assignment and back check rules nobody has written down. Cutting over mid round is how organisations lose data.
Who maintains this once the grant that funded it closes?
Somebody has to be paid to, and this is the question that decides whether the build survives. Budget fifteen to twenty percent of build cost annually for hosting, support and continued development, and name the internal owner before you start. Organisations that fund from unrestricted reserves, a multi year framework or a consortium arrangement keep their platforms running. Those that fund from a single project grant tend to abandon them within two cycles.
How do we protect household data collected in the field?
Separate identifiers from analytical data at rest and restrict the identified layer to named roles with access logging. Make de-identified export the default so an analyst wanting a quick file cannot create an incident. Devices should carry only the assigned caseload, encrypted, with remote revocation that has actually been tested. Record consent as a structured fact including the language used and the version of the consent text, with automatic disposal at retention.
Who builds this kind of platform for humanitarian organisations?
Specialist civic technology outfits and custom development firms with field operations experience. Digital Heroes suits agencies that need contracting and intellectual property assignment in their own jurisdiction, operating as an India LLP, a US LLC and a UK LTD, and that want the quality rules and indicator definitions written down before code starts. Over 2,000 projects delivered, a fifty plus person team, and Fiverr Vetted Pro status back that up.
What makes Digital Heroes different from a generic dev shop for this problem?
The refusal to rebuild what already works. The standard recommendation here is to keep KoboToolbox or SurveyCTO for capture and build only the workforce, quality and indicator layers, which removes about a third of the budget and the riskiest engineering. Alongside that, the client owns the repository from the first commit, so a platform funded by one grant can be handed to whoever runs the programme next.
How do we verify a development partner is legitimate before transferring funds?
Confirm the D-U-N-S registration matches the legal entity named on your contract, then read the public Clutch and Trustpilot profiles for reviews from comparable work rather than the headline rating. Ask which entity signs and under which law, since jurisdiction decides your recourse. Request a redacted prior contract showing intellectual property assignment, and require repository and cloud account access from week one rather than at handover.
Should I launch with an MVP or wait until the app feels complete?
Launch the minimum viable product, because no app is ever complete and real store reviews reshape a roadmap faster than any internal debate. In Digital Heroes delivery experience, a focused first release with five to eight core features runs 40 to 60% less than the founder's full wish list and ships months sooner. The discipline is choosing the one job the app must do perfectly and deferring everything else to updates.
What changes when my app grows from 1,000 to 100,000 users?
Scaling from 1,000 to 100,000 users mostly changes the backend and the bills, not the app on the phone. Expect database tuning, caching, and a move off entry-level hosting tiers, with infrastructure costs climbing from tens of dollars a month into the hundreds or low thousands. This is also where no-code backends hit hard ceilings, Bubble's workload unit pricing being the classic example, which is why products expecting real scale either start custom or plan the migration early.
Who owns the source code when an agency builds my app?
You should own the source code outright, and the contract must say it plainly with an intellectual property assignment that transfers ownership on final payment. Watch for agreements that only license the code to you, keep it in the agency's repository, or register the Apple and Google developer accounts under the agency's name. Insist on code delivered into a repository you control from week one, not at final handover.
What security does my app need if it takes payments?
Never store card numbers yourself: run payments through Stripe, Braintree, or a similar processor's software development kit so the heaviest compliance burden stays with the processor. Beyond that, a properly built app encrypts all traffic, keeps session tokens in the platform's secure storage (iOS Keychain, Android Keystore), and enforces backend rules so one user can never read another's records. Ask a prospective agency how they handle those three things; vague answers are disqualifying.
How long until a business app pays for itself?
Internal and operations apps pay back fastest, typically inside 12 to 24 months across Digital Heroes projects, because the savings are countable: hours of manual entry removed, errors avoided, jobs scheduled tighter. Consumer apps are slower and riskier because payback depends on acquisition costs you only partly control. Before building, write down the one number the app must move, bookings per week or support calls per day, and have the agency design around it.
Can I build my product on a no-code tool like Bubble instead of hiring developers?
For testing whether anyone wants the product, yes, and Bubble's paid plans start at $29 a month, which is the cheapest validation you will ever buy. The ceiling arrives with complex data relationships, heavy integrations, performance at a few thousand users, and the fact that you cannot export a Bubble app to servers you control. A path many Digital Heroes clients take: prove demand on no-code, then rebuild custom once revenue justifies it, treating the no-code version as a paid prototype rather than a foundation.
We run everything on spreadsheets and Airtable. How do we know it's time for custom software?
The reliable signals are re-typing the same data into multiple tools, one employee acting as human middleware between systems, and errors appearing in handoffs between teams. Hard limits force the issue too: Airtable's Team plan caps at 50,000 records per base, and Business costs $45 per seat per month, so a 20-person team pays about $10,800 a year for a tool it has already outgrown. When workarounds consume more hours than the tools save, the spreadsheet era is over.
Does my app need to be HIPAA or GDPR compliant?
HIPAA applies if the app handles US health information for providers, insurers, or their vendors; GDPR applies the moment you have users in the EU, wherever your company is based. Both reshape the build: HIPAA requires hosting vendors that will sign a business associate agreement, and GDPR requires consent, data export, and account deletion flows. No-code platforms generally will not sign a business associate agreement on standard plans, which by itself pushes most health apps to custom development.
Should I hire a freelancer or an agency to build my app?
A strong freelancer suits a small, tightly defined app where you supply the product direction and design references yourself; in the competing quotes Digital Heroes sees, freelance rates usually run $30 to $100 an hour. An agency earns its overhead when you need design, mobile, backend, and testing in one accountable team, and when the project cannot stall because one person disappears. A rough dividing line is $25,000 of scope: below it, a good freelancer is often the better buy.
How many people does it actually take to build a mobile app?
A typical agency team is four to six people: a project lead, a designer, one or two mobile developers, a backend developer, and a tester, most of them part-time on your project. A lean first version can ship with three. Be skeptical of one person claiming to cover design, mobile, backend, and testing alone on a complex app; something on that list is being skipped, and it is usually testing.
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.
Can a custom app integrate with the software my business already runs?
A custom app can connect to almost anything your business already runs, which is one of the main reasons buyers outgrow no-code builders. Custom code can talk to anything with an application programming interface, including QuickBooks, Salesforce, Shopify, Stripe, and your internal databases, while app builders restrict you to their catalog of prebuilt connectors. List every system the app must touch before requesting quotes; integrations move the price more than screen count does.
Who can build a custom mobile app system?
Digital Heroes builds custom mobile app 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 mobile app 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
Published · Last updated .
