Building it costs $150K to $340K.Embedding it costs two tickets.
Customer-facing analytics takes 6 to 12 months and 1.5 to 2 engineers to build, and the maintenance never ends. Here's the math with its inputs shown, a worksheet you can rerun with your own numbers, and what the other path costs.
The estimate everyone approves is the first row. The rows under it are what the line actually contains.
It starts as one line on the roadmap
"Customer dashboards, Q2." A charting library, a few queries, maybe three sprints. Your customers have asked for it, and your engineering lead has already gone quiet in the meeting where it came up.
The first real requirement arrives
Every customer sees only their own data. That constraint reaches the schema, the API and every screen, and it has to hold on paths nobody thought to test.
Configuration becomes a feature
The second customer wants a different view from the first. Now somebody builds the system that lets them have one, and documents it, and supports it.
Then exports, scheduled reports, permissions
Each one obvious in hindsight. None of them in the sprint that got approved, and each arriving as a reasonable request you can't refuse.
Two roadmap items slide to Q3
The features queued behind this one move, and the engineers who would have shipped them are here instead, on a dashboard project.
That's how one line becomes a second product, staffed by the engineers who were supposed to be building your first one.
The build costs $150K to $340K.Here's every line of it.
No hand-waving, and no scary figure without its inputs. Stack Overflow's 2024 Developer Survey puts the median US back-end developer at roughly $170,000. Standard fully-loaded multipliers for benefits, payroll taxes, equity and tooling run 1.4x to 1.6x, so one senior engineer costs $240K to $280K a year. Everything below comes from that.
Run it with your own numbers
Four inputs. The output is your year-one engineering cost, not ours.
Engineering time only. Add your own numbers for design, QA, and the security review if you want the full picture.
Where the published band comes from
A credible first version needs about 1.5 to 2 engineers, backend plus part-time frontend. The phases overlap, which is why the calendar is 6 to 12 months and not the sum of the rows.
| Work | Time | Cost |
|---|---|---|
| Visualization layer: charts, filters, layouts | 1 to 2 months | ~$60K |
| Multi-tenancy, row-level security, hardening | 4 to 6 months | $90K to $160K |
| AI query layer, if your customers expect one | 4 to 6 months | $80K to $120K |
| First-year build total | 6 to 12 months | $150K to $340K |
Year one is the cheap year.Then the maintenance starts.
Plan on at least 20% of an engineer, permanently: schema changes, broken charts, tenant edge cases, performance. If there's an AI layer, add answer evaluation plus keeping the metric definitions it relies on current. Teams that have done this land in the $470K to $740K range over three years, once the enterprise tenancy rework arrives.
Maintenance is derived from the same inputs: 20% of one fully-loaded engineer at $240K to $280K a year. The rework is the swing factor, which is why the three-year range is wide.
Your estimate covers building it.The hard part is keeping it true.
Anthropic's own internal analytics agent drifted from ~95% accuracy at launch to ~65% in a month, because documentation went stale as their data model changed underneath it (Anthropic). Price that into the build. It means someone re-verifying every customer-facing answer, every sprint, for as long as the product exists, and that person is nowhere on the estimate you were handed. Page 2 walks the failure step by step. Here, it's three questions your estimate has to answer before anyone signs it.
Multi-tenancy
Can every customer see only their own data, enforced at the data layer and tested on every query path, including the ones the AI writes itself? Close isn't a passing grade here, and it's the phase that takes 4 to 6 of your months.
Whose definition wins
When two teams define "active customer" differently, which definition does a customer-facing answer use? Someone has to own that dictionary and keep it current as the product changes. Name them before you start.
The fast follows
Exports, scheduled reports, permissions, sharing, mobile. Which of these are inside the estimate, and which arrive afterward as fast follows that never stop following? Every one of them is a sprint your first product doesn't get.
Both paths, on the cost questions
The build column is populated from the worksheet above, at its default inputs.
Sometimes it is. Build if:
Analytics is the product
Not a feature of it. If dashboards are what you sell, own them. The maintenance tax is your core competency, and you should be better at it than any vendor.
Nothing underneath changes
Simple, single-tenant, and nearly static. The maintenance tax shrinks when nothing changes underneath the answers, and drift stops being the standing risk.
Your data team has spare room
A real data platform team, with real spare capacity, whose core competency is reporting. Not a team that's already behind on the roadmap you just added to.
2 tickets from your engineers.Analytics under your brand.
Everything the build path would have put on your roadmap for the next year already runs inside Querri. Your brand, your UI, powered by Querri underneath.
The easiest yes your engineering team will give you
Provision data access, drop in the embed SDK. Your instance is live with your data in about a week, and customer-facing after tuning, typically two to four weeks depending on how fast you want to move. The tuning is done by your product people, so it never queues behind an engineering sprint.
Two tickets, then done
The engineering ask ends at provisioning data access and dropping in the SDK. No config system, no permission project, no fast-follow queue.
The 20% engineer stays yours
Schema changes, broken charts, tenant edge cases, performance. The maintenance line you would have staffed forever is on Querri's side of the ledger.
Your roadmap back
The two features that would have slipped behind the dashboard project ship instead. That's the number your board actually asked about.
Your brand
Ships in your UI under your name. Your sales team demos it as yours, because it is.
Every one of these is a line on the estimate
Your customers are already asking these. Each one is a view to build, a permission rule to enforce, and a chart that has to stay right every time the schema moves.
The short answers
How long does it take to go live?
Two gates. Your white-labeled instance is up with your data connected in about a week. Customer-facing launch typically lands two to four weeks in, after tuning, and the tuning is done by your product people, not your engineers.
What does our engineering team actually have to do?
Two tickets: provision data access, and drop in the embed SDK. Everything after that, including views, instructions, and refinement, happens without an engineering sprint.
What does it cost?
Less than the cost of one engineer, on consumption pricing. Viewing and sharing are free. Compare that with the $240K to $280K fully-loaded cost of each engineer on the build path.
We already started building. Is it too late to switch?
No, and the assessment is a cheap way to decide. If your data foundation is solid, you'll know. If it isn't, better to learn that from a 4 to 5 page report than from a missed quarter.
Can't we just use a charting library?
For the charts, yes. The expensive parts are everything underneath: multi-tenancy, security, the data model, and keeping answers accurate as your schema changes. That's where the 6 to 12 months go.
One argument, four doors
Every page shows its numbers and its sources, including the honest cases where building is the right call.
Everyone is asking when you ship AI
Anthropic's own analytics agent went from 21% to 95%+ answer accuracy on the same model. The model was never the hard part.
Your export button is a demand signal
Analytics rated 43% of an application's perceived value a decade ago. Your customers want it more now, not less.
The premium tier you haven't priced yet
A median 25% price premium for analytics, set on static dashboards a decade ago. Build vs. buy decides when you start collecting.
You know what the build costs.See what your data supports.
Start with one export, or challenge us with as many sources and connectors as you'd like. In 3 to 5 business days you get a branded 4 to 5 page report: what's clean, what's broken, and the questions your customers could already be asking your product. Free, and yours to keep.