On Page Navigation
Business intelligence on infrastructure you own. We connect your databases, build the queries and dashboards your business actually runs on, and train your team to build the rest. One flat monthly price.

How we approach it
Most reporting engagements end the week the dashboards are handed over, which is roughly when the business changes and the numbers start drifting. We work the other way round. The build matters, and so does month six, when somebody asks a question the dashboard was never designed to answer. Where the data underneath comes from an ERP we also run, that continuity gets easier again, and our managed ERPNext work is how a good number of clients arrive here.
The build itself starts by agreeing which numbers a decision actually turns on, because a dashboard with forty tiles is a dashboard nobody reads. Definitions get written down, so revenue means the same thing in two places and nobody argues about whose figure is right. Queries run against the source rather than an export that goes stale. Access is set so people see their own numbers without seeing everybody's. And the arrangement carries on past handover, which is when the questions that matter usually start arriving.
Not with the database. The first sessions are about the decisions you are trying to make and the numbers you already argue about in meetings. Modelling the data is the easy half. Knowing which twelve numbers matter, out of the several hundred your systems could produce, is the half that decides whether anybody opens the dashboard twice.
Revenue means something slightly different to your finance lead and your sales lead, and each is correct inside their own frame. A report cannot hold both. We make you settle every contested measure before it goes in, record which definition won, and put that document somewhere findable. Least technical part of the work, largest number of arguments prevented.
Insights can inherit your existing Frappe role permissions, but only for site data sources. Point it at an external MySQL or PostgreSQL database and that inheritance quietly does not apply, with nothing anywhere to say so. We establish which of your sources sit on which side of that line before designing access, rather than after somebody reaches a table they should not have.
The database user we connect with cannot write. Two minute decision at setup, and it removes a whole category of incident, because no query anybody builds later can damage the system it is reading. We also agree which sources are safe to query directly and which want a replica. A heavy report competing with your production application is a performance problem wearing a reporting problem as a disguise.
The point of a query builder is that people stop queuing for reports. That only happens if two or three people on your side learn it properly, so we train those people rather than running a session for everyone. Pick them for curiosity rather than seniority. Whoever already rebuilds the same spreadsheet every month is usually the right choice.
Reporting accumulates. Nobody ever volunteers to delete a chart, so dashboards grow until the number that matters is somewhere on page two. Once a quarter we look at what is actually being opened and propose removals. Expect us to argue for fewer dashboards than you asked for.
Subscription against a one off build
A reporting project hands you the answers that were true on the day it finished. Businesses add product lines, restructure teams and change what they count, and every one of those events puts a dashboard slightly out of date without anyone noticing. A subscription absorbs that. The same reasoning runs across the rest of the platform, which is why we support each of the official Frappe applications in our catalogue on the same footing.


A project writes down what revenue means once. Six months later a new product line arrives, nobody revisits the definition, and two teams start quoting different figures off the same dashboard. We revisit the contested measures on a schedule instead of waiting for somebody to complain.
New tables in a connected database do not appear in Insights until the table list is refreshed by hand. On a project that step lives in a handover document nobody opens. Here it lives in a runbook with a name against it, which is the difference between a missing table and a missing afternoon.
Teams change and people move, so a permission model built in month one describes a company that no longer exists by month twelve. Nobody raises a ticket about having too much access. We review it whether or not anyone asks.
Fiscal calendars move, and a date table built once around last year's year end starts quietly misreporting the current one. Small artefact, a lot of weight on it, since every period comparison on the site joins against it.
A build only ever adds. With nobody whose job is to remove things you end up with forty dashboards and three that matter, and a new starter cannot tell which is which. We propose removals each quarter and you decide.
A project has no reason to mention that your data has outgrown the platform, or that the source you want to add next is one Insights cannot read. We do, because we are still here in month eighteen and by then it is our problem too.
Who this suits
Insights earns its place when the data already exists and the difficulty is reaching it. If you are a small team whose numbers still live mostly in one accounting package and a spreadsheet, this is more platform than the question deserves, and managed reporting on Zoho Analytics will get you there faster and for less. The four situations below are where the heavier tool starts paying for itself.




What the first weeks look like
Two to five weeks for most rollouts. Connections and charts are quick work. What sets the pace is your side agreeing what the numbers mean, and that conversation is worth taking properly rather than rushing.
1
Not from a requirements document. From your real spreadsheets, the exports somebody pivots by hand every month, and the dashboard in whatever you are leaving. What a business genuinely needs is visible in what it keeps rebuilding, and the list is usually shorter than anyone expects.
2
Data sources, the date table on your fiscal calendar, the joins your reporting depends on, then the queries and dashboards on top. Every headline number gets reconciled against its source before anybody sees it. A dashboard that disagrees with the accounts once will not be trusted again, and rebuilding that trust costs more than the original build.
3
Training for the two or three people who will build their own reports from here, working on your data rather than a demonstration set. Then access control, the written record of what each measure means, and the request queue for everything that comes afterwards.
Pricing
Standalone Service
DETAILS
Standalone Service
DETAILS
Testimonials
Everything we set up and support



















