Service · build
Process Rebuild
For the work that crosses systems which don’t talk to each other. We design it first and sell the design on its own, so you can see exactly what you’d be committing to before you commit to it.
When this is the right answer
Four tests. Any one of them and it’s a rebuild
Most portal problems are fixable inside HubSpot. These aren’t, and pretending otherwise is how fixed-price engagements go wrong for both sides.
Test one
It crosses a boundary
Data has to move between HubSpot and your ERP, billing, product or warehouse — or two systems have to agree what a record means.
Test two
It needs a parallel run
The old way and the new way have to run side by side while you verify. That’s never a two-week job, whatever anyone tells you.
Test three
Someone else builds part of it
Your developer, your vendor, or your IT team. As soon as delivery depends on a third party’s timeline, fixed price stops being honest.
Test four
Failure is silent
A broken workflow shows up the same day. A broken sync shows up as wrong invoices six weeks later. Those need different safeguards.
Design first
You buy the design before you buy the build
$6,500, sold on its own, with no obligation to continue.
- Current-state map of how the process actually runs today, including the parts people do manually and don’t mention.
- Target-state design: what moves where, which system owns which field, and what happens when a record fails.
- A costed build estimate with the assumptions written down, so you can see what would make it more expensive.
- A parallel-run and cutover plan, including how you’d back out if it goes wrong.
You can take the design to another partner, run it internally, or decide not to do it at all. Roughly one in four designs we sell doesn’t become a build, and that’s a legitimate outcome rather than a lost sale.
The build
Priced on what it’s worth, not on how long we sat there
Pricing
A share of year-one value
Quoted as a single fixed number derived from the design. We don’t publish a floor, because a number without a scope is meaningless and invites the wrong conversation.
Timeline
Ten to fourteen weeks
Including the parallel run. We don’t compress this, and we’d rather turn work down than agree to a date that requires skipping verification.
Cutover
Parallel run is mandatory
Both processes run together until the new one matches the old one on real records. No exceptions, including when the old process is obviously worse.
Fair warning
Reasons we’ve declined this work
- The real problem was a decision nobody had made yet. No integration fixes an unresolved argument about what a stage means.
- The other system was being replaced within the year. Build it twice or wait — usually wait.
- The value didn’t justify it. Some processes are annoying and cheap to leave alone.
- Nobody internally owned it. A rebuild with no owner degrades faster than the thing it replaced.
“They went above and beyond to create seamless integrations and automations that will save us time and money. They solved a problem for us that we couldn’t solve ourselves.”Vanessa ChengCo-Founder · OrbViz · Technology
Not sure whether it’s a rebuild or a fix
Most of the time it isn’t, and we’ll tell you that for free. Start with a Leak Check and we’ll show you where the boundary actually sits in your portal.