Mortgage Secondary Market Trading Software: Custom Build or Optimal Blue
Buy. Keep Optimal Blue or Polly for rate sheets and pricing, and keep a hedge advisory relationship if you originate under about $200 million a month.
On this page
Buy. Keep Optimal Blue or Polly for rate sheets and pricing, and keep a hedge advisory relationship if you originate under about $200 million a month. Build only the thin decision layer on top: your own pull through model, best execution across the outlets you are actually approved for, and a lock desk exception ledger. Full desk platforms suit very few lenders.
What Optimal Blue, MCT, Polly and ICE Compass actually do well
Optimal Blue is the widest product and pricing engine in the market and it should generally stay in your stack. Ingesting investor pricing, maintaining loan level price adjustments and generating rate sheets correct at 8am is a maintenance treadmill that buys you nothing a borrower can see. Polly does the same job with a cleaner interface surface, which matters a great deal if you intend to build anything on top. MCT gives you a hedge advisory relationship plus a platform, and for a desk under a few hundred million a month that combination is usually cheaper and safer than hiring the equivalent expertise. ICE Compass Analytics is serious analytics for larger desks.
Since this page is on a development firm's site, the useful thing to say first is the thing against our interest. Most independent mortgage banks should not build a secondary marketing system. If you originate under roughly $100 million a month in conventional agency product, sell best efforts, and do not retain servicing, your interest rate risk is small, your delivery outlets are few, and an advisory relationship covers you for a fraction of a build. Rebuilding a pricing engine there is a hobby with a budget line.
Buy, and stop reading, if most of these hold:
- You sell best efforts, so fallout is the investor's problem rather than yours.
- You originate one broad product set that every third party engine prices properly.
- You do not retain mortgage servicing rights and have no plan to.
- Your locked pipeline rarely exceeds a couple of hundred million.
- Nobody at your shop is being asked to defend a hedge result to a board.
Where they stop: pull through, approvals and the exception ledger
Every one of those products stops at the same place, and it is not a product weakness. It is that three inputs are yours and cannot be theirs.
The first is pull through. Most desks carry it as a handful of buckets by lock stage, set from a rough historical average and reviewed occasionally. Those numbers are wrong in exactly the moment they matter. Pull through is driven above all by rate incentive: when market rates fall fifty basis points below the note rate on a locked loan, that borrower has a reason to renegotiate or walk, and your fallout rises precisely while your forward position is losing. It also varies by channel, purpose, lock term, remaining days and whether an appraisal is in hand. A blended number blends away every one of those signals, and no vendor will let you inspect or retrain their model on your own funded history.
The second is best execution. A loan can go to the agency cash window, into a pool, out as a correspondent whole loan sale, through an assignment of trade, or onto your own balance sheet. Real comparison has to price the whole loan including servicing: a service release premium today against a servicing asset whose value depends on prepayment expectations, guarantee fee buy up and buy down grids that change the coupon and therefore the pool, and specified pool payups for low loan balance, geographic or high loan-to-value stories that are real money and routinely ignored. Vendor best execution knows public investor grids. It does not know your investor approvals, your delivery capacity this month or your appetite for retained servicing this quarter.
The third is the least glamorous and it is frequently the one that pays for the project. Extensions, relocks, renegotiations, float downs and worst case pricing exceptions get granted every day, usually on a phone call with a producer. Each has a cost and almost nobody attributes it back to the loan, the branch or the loan officer who asked. Those exceptions live in your loan origination system and your email, so no third party platform holds them. In most shops a small number of producers consume a large share of the exception budget and nobody has ever been able to prove it.
The arithmetic: basis points per locked loan versus the cost to build
Price both sides in basis points of locked volume, because that is the only unit that makes them comparable.
Add your annual pricing engine subscription, your hedge advisory fee, any per loan charges, and the fully loaded cost of the secondary analyst who rebuilds the position report in Excel every morning. Divide by annual locked volume. Most lenders who run this land in the low single digits of basis points, but run yours, because the analyst's time is the line everyone leaves out and often the largest.
Now the build side. A decision layer sitting on top of a bought pricing engine runs $70,000 to $150,000. Take the midpoint, amortise across five years, add year two support, and call it $30,000 to $50,000 a year. At $100 million a month of locks, that is roughly a quarter of a basis point. At $500 million a month it is well under a tenth.
Which tells you that subscription arithmetic is not the argument. The argument is the gap between the lock and the hedge. A desk running $300 million a month carries roughly $600 million of open commitments at any moment. A forty basis point move against an unhedged sliver of that is not a rounding error, it is the origination margin for the month. The position report that drove this morning's trade already describes a pipeline that no longer exists, because locks are events and the hedge is a batch joined by a person with a spreadsheet.
So the crossover, stated plainly: around $200 million a month in locked volume, or roughly 600 locks a month, combined with retained servicing or a serious decision about it. Below that line, buy and hire an advisor. Above it, the thin layer is worth building and the licence savings are irrelevant to the case.
What a custom desk costs, and whether it is worth building
A decision layer over your existing pricing engine runs $70,000 to $150,000 and ships in 10 to 16 weeks in our delivery experience. That covers an event driven position that updates on every lock, change, cancellation, fallout and funding, a loan level pull through model trained on your own history with rate incentive recomputed against live pricing, best execution across only the outlets you are approved for, and the exception ledger. Your engine keeps producing rate sheets.
A full desk platform adding trade capture with pair offs, margin call tracking against your broker dealer counterparties, statement reconciliation, investor commitment management and daily profit and loss attribution runs $200,000 to $500,000 phased over 8 to 14 months.
Then the two lines nobody quotes.
- Data preparation and migration: 10 to 25 percent of the build. Here that means your lock history. If your loan origination system never recorded the market level at lock time, the model cannot learn rate incentive, and reconstructing it from archived rate sheets is real weeks of work before any modelling begins. Assume 18 to 24 months of clean lock and funding history is the minimum input.
- Year two: 15 to 20 percent of build cost annually. Model retraining as your channel mix shifts, each new broker dealer statement format, each new investor commitment structure, and the accounting support your auditor will want when they ask how a fair value mark on a rate lock commitment was derived.
What pushes the number up: the count of investors and delivery outlets, since each commitment structure is its own modelling problem; retained servicing, which adds a genuine valuation workstream; the number of counterparties you clear with; and whether you pull locks from Encompass in real time or from a nightly file, which is a several week difference on its own.
Four situations where building beats buying
Custom versus off the shelf here comes down to four conditions.
- Regulatory and accounting fit. Rate lock commitments and forward sales are carried at fair value, and your auditor will ask how each mark was derived. If the answer involves a spreadsheet nobody can reproduce, you have an audit problem that a vendor report does not solve, because the vendor cannot evidence your inputs.
- Scale economics. Past roughly $200 million a month, the latency between the lock event and the hedge adjustment costs more in a single volatile week than the whole build. That is the entire case and it does not need embellishing.
- A workflow that is your competitive advantage. If you originate non-QM, jumbo held on balance sheet, construction to permanent, or you are a credit union pricing members differently from the market, no third party engine prices your product properly. Your execution logic is the business, and renting a generic version of it caps what you can originate.
- Integration sprawl across three or more systems. Loan origination system, pricing engine, broker dealer statements, servicing valuation, general ledger. When the position lives in an export, the export is the lag and the lag is the loss.
How to decide in a week
Run this and you will know by Friday. Monday, export every lock from the last twelve months and mark which funded. Tuesday, check whether the market level at lock was recorded on each one. Wednesday, compute realised pull through by rate incentive bucket, splitting loans that ended up in the money against those that did not. Thursday, compare that curve against the bucket assumptions your desk actually used. Friday, total every extension, relock and renegotiation granted in those twelve months and attribute each to a branch.
Two outcomes. If Tuesday stops you because the market level at lock was never stored, that is your answer and your first project, and it is a data capture job rather than a modelling one. If Wednesday shows realised pull through swinging widely across incentive buckets while your desk used one number, you have quantified the gap in a week for the cost of an analyst's time.
Then buy a paid discovery phase rather than a build. At Digital Heroes that produces a signed product requirements document before any code, covering the data model, the model validation approach, permissions and acceptance criteria. You own it whether we build or not, and it is the only way to get quotes describing the same system.
We are the wrong firm for you if you want a body shop billing hourly, if you need a hedge advisor rather than a builder, or if you originate under $100 million a month best efforts. We are more than fifty specialists across India LLP, US LLC and UK LTD entities, so your intellectual property assigns under your own law, and you meet the named team before signing rather than a bench in month two. We run our own products, ShopScore, HeroCheckout and Section Vault, so the people choosing your architecture live with those decisions on their own revenue. Verify the rest on Clutch, Trustpilot, Fiverr Vetted Pro and our D-U-N-S record.
Book a 30-minute call with Digital Heroes and get a written plan and a fixed quote within 48 hours.
The evidence behind this guide
Independent findings on why this investment pays off. Every link goes to the primary source.
- The average developer spends more than 17 hours a week dealing with maintenance issues such as debugging and refactoring, and about four of those hours on 'bad code' - waste that equates to nearly $85 billion annually worldwide in opportunity cost. Source: Stripe (2018) →
- Analyst estimates place CRM implementation failure rates broadly between roughly 30% and 70% (Johnny Grow cites Forrester at 47%), with low user adoption repeatedly cited as a leading cause of failed CRM projects (this being Johnny Grow's own analysis, not a Forrester attribution). Source: Johnny Grow (industry analysis citing Gartner/Forrester) (2025) →
- The right combination of digital transformation actions can unlock as much as US$1.25 trillion in additional market capitalization across Fortune 500 companies, while the wrong combinations put more than US$1.5 trillion at risk; companies with all three core factors (strategy, aligned technology, and change capability) saw a 5% market-value lift relative to peers. Source: Deloitte (2023) →
- In the Flexera 2025 State of ITAM report, respondents reported roughly 33% of SaaS spend is wasted, underscoring how paying for off-the-shelf seats and tiers that go unused erodes the supposed cost advantage of generic SaaS. Source: Flexera (2025) →
Frequently asked questions
How much does it cost to build custom secondary marketing software?
A decision layer over a bought pricing engine, covering an event driven position, a pull through model on your own history, best execution and an exception ledger, runs $70,000 to $150,000 in 10 to 16 weeks. A full desk platform with trade capture, margin calls and profit and loss attribution runs $200,000 to $500,000 across 8 to 14 months. Budget another 10 to 25 percent for historical data preparation.
How long before a custom pull through model is trustworthy enough to trade on?
Plan on running it alongside your existing assumptions for a full quarter before it drives a trade. Validation means holding out a period of your own funded history, backtesting the coverage the model would have recommended against what actually funded, and showing error by rate incentive bucket. A desk that switches on day one has skipped the only step that tells you whether the model works in a moving market.
Who owns the trained model and the funded loan history?
You should own all of it, and settle this in writing before kickoff. Your funded loan history is the most valuable asset in the project, because it is the only data from which your own pull through behaviour can be learned. Insist on the repository, the cloud accounts, the training data and the model weights sitting in infrastructure you control. At Digital Heroes the client owns everything from the first commit.
What happens if we switch pricing engines after building the decision layer?
It should cost weeks, not a rebuild, provided the layer was designed against an interface boundary rather than against one vendor's data shapes. Pricing, locks and investor grids arrive through adapters, so replacing Optimal Blue with Polly or the reverse means writing one adapter. Ask any developer to show you that boundary in the design before you sign, because the alternative quietly makes your vendor unswitchable.
Can we build only the lock desk exception ledger?
Yes, and it is the cheapest useful thing in this category. Every extension, relock, renegotiation and float down becomes a ledger entry with an amount, an approver, a reason and a link to the loan, and the monthly margin report then shows realised margin after exceptions by branch and originator. It touches no trading logic, carries almost no risk, and usually starts an uncomfortable and valuable conversation within one month.
What is the difference between best efforts and mandatory delivery for software purposes?
Under best efforts you deliver the loan if it closes and the investor carries fallout risk, so your software mostly needs to track commitments. Under mandatory delivery you owe the loan regardless, fallout becomes your exposure, and pair offs cost real money, which is why pull through modelling, hedge coverage and trade capture only become necessary once you go mandatory. A developer who cannot explain this will model the wrong system.
Should a credit union or bank build rather than buy here?
More often than an independent mortgage bank of the same size, yes. If your best execution includes a portfolio retention decision, or you price members differently from the market, no third party engine models the outlet that matters most to you. The build is not about hedging efficiency in that case, it is about comparing a balance sheet decision against a sale on the same page, which vendors do not attempt.
What happens if the market moves hard while our position report is an export?
You find out how large the gap was afterwards. A desk with $600 million of open commitments and a morning batch position is unhedged against every lock, change and fallout that arrived since the extract ran. On a quiet week nobody notices. On a jobs number that moves the ten year forty basis points, the difference between an event driven position and a morning export is the month's origination margin.
Can artificial intelligence forecast our pull through better than a model on our own data?
Not meaningfully, and be careful with anyone claiming otherwise. Pull through is driven by a small number of observable factors, above all rate incentive, and a well specified model trained on eighteen to twenty four months of your own locks captures them. What matters is that the model is yours, inspectable and retrainable as your channel mix shifts. Sophistication of technique is not the constraint here, data quality is.
How do we compare quotes when every firm scoped something different?
Pay for discovery first and take the written specification to everyone on your shortlist. It should name the data model, the outlets in scope, the validation method for the pull through model, the accounting evidence required and the acceptance criteria. Without it you will receive four quotes that priced four different systems and the cheapest will be the one that understood the least. Any firm refusing to quote against it has told you something.
Can a custom internal tool connect to QuickBooks, Salesforce, and the other software we already use?
Yes, and integrations are usually the strongest argument for going custom instead of chaining tools together with Zapier. QuickBooks, Salesforce, Shopify, Stripe, Slack, and Google Workspace all have mature APIs, and each integration typically adds $1,500 to $5,000 to a Digital Heroes build depending on how much two-way syncing you need. The honest caveat is legacy industry software without an API, which may need file-based imports instead of a live connection, so list every system in the first conversation.
Will a custom internal tool scale as our company grows?
Yes, provided it sits on a standard stack with a real database: PostgreSQL comfortably handles millions of records, and adding users costs hosting pennies rather than per-seat fees. The real scaling risks are organizational, not technical: new departments want features, processes change, and the tool needs a budget line to evolve. Set aside a small quarterly improvement budget instead of treating launch as the finish line, and the tool stays useful for a decade rather than getting rebuilt every two years.
Is a freelancer or an agency better for building an internal tool?
A solid freelancer works for a single-workflow tool under roughly $10,000, if you accept that one person holds all the knowledge. An agency earns its premium once the tool spans departments or integrations, because you get a developer, a designer, and a project manager plus continuity when someone leaves or gets sick. The hidden freelancer cost appears 18 months later when you need changes and the original builder has moved on, a rescue situation Digital Heroes is hired for regularly.
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.
How do I know when spreadsheets are no longer enough to run my operations?
Replace the spreadsheet once more than three people edit it, versions travel by email, or a single broken formula could cost real money. Other reliable signals: staff keep personal shadow copies, month-end reporting takes days of manual assembly, and nobody can say who changed a number or why. In Digital Heroes discovery calls the tipping point is almost always a specific expensive error, a mispriced quote, a missed order, or payroll built on a tab someone sorted wrong.
How many SaaS seats do we need before building custom becomes cheaper?
The crossover usually shows up between 20 and 50 seats on premium tiers. Salesforce Enterprise lists at $165 per user per month, so 40 users cost about $79,000 a year in subscriptions, which is real money against a custom system you would own outright. Run the comparison over three years: if subscription spend beats the build cost plus 15-20% annual maintenance, custom wins on price before you even count workflow fit.
How long does it take to build an internal tool from scratch?
A working first version typically ships in 4 to 8 weeks, and larger multi-module tools run 10 to 16 weeks. Across Digital Heroes internal tool projects the schedule splits into roughly one week of process mapping, 3 to 6 weeks of build, and 1 to 2 weeks of testing with your actual staff. The most common delay is not development but waiting on the client for sample data and workflow decisions, so name one internal owner before kickoff.
Is a custom internal tool secure enough for HR records and financial data?
A properly built custom tool is generally safer for sensitive data than the shared spreadsheet it replaces, because you get role-based access, audit logs, encrypted storage, and the ability to cut one person's access instantly. Ask the agency specifically for encryption in transit and at rest, permissions down to the field level, and an audit trail showing who viewed or changed each record. If HIPAA, GDPR, or SOC 2 expectations from enterprise clients apply to you, raise it before the quote, because compliance features add real scope.
Who can build a custom internal tools system?
Digital Heroes builds custom internal tools 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 internal tools 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 Priya.
Senior PR & Comms Manager · New-york · 13 posts on the site
Published · Last updated .
