Finance & FP&A

How to run monthly flux and variance analysis as Small Investor Relations Teams

Finance & FP&AFor Small Investor Relations Teams3 apps12 steps~24 min to set up

Your two-person IR team closes the books each month and then spends the next week manually reconciling actuals against what you told LPs in the last quarterly letter. You're pulling QuickBooks exports into Excel, cross-referencing Plaid transactions to catch anything the bookkeeper miscategorized, and trying to explain why management fees came in 12% below forecast when the GP wants an answer in 90 minutes. The institutional IR platforms assume you have a dedicated FP&A analyst to do this. You don't. You have a pivot table, a shared Google Sheet that someone keeps breaking, and a standing Wednesday meeting where half the agenda is 'why does this number not match that number.'

Build this on Starch → See the apps used

Finance & FP&AFor Small Investor Relations Teams3 apps12 steps~24 min to set up

Outcome

What you'll set up

A live flux and variance dashboard that pulls from QuickBooks actuals and Plaid cash movements on a schedule, so your actuals are never more than 24 hours stale when you sit down for month-end

An automated variance narrative generator that flags every account where actuals deviated more than a defined threshold from prior period or budget, and drafts plain-English explanations your GP or CFO can read without a translation

A monthly close checklist automation that pings you in Slack when data is refreshed and ready, reducing the 'is this the right number?' back-and-forth that eats Tuesday mornings

The Starch recipe

Apps, data, and prompts

The combination of Starch apps, the data sources they pull from, and the prompts you use to drive them.

Apps used

Runway Analysis Budgeting Transaction Insights

Data sources & config

Starch syncs your QuickBooks data on a schedule — invoices, bills, payments, vendors, journal entries — and syncs your Plaid bank transactions and balances on the same cadence. The Runway Analysis and Transaction Insights starter apps are wired to these connections out of the box; the Budgeting app adds budget-vs-actual comparison layers on top. For LP portals like Juniper Square or iLevel that don't have a direct integration, Starch automates them through your browser — no API needed.

Prompts to copy

Build me a monthly flux analysis dashboard that pulls actuals from QuickBooks and Plaid, shows every P&L line item side-by-side with last month and last year, calculates the dollar and percent variance, and flags any line where variance exceeds 10% or $5,000 — whichever is smaller

For each flagged variance, draft a one-sentence plain-English explanation I can paste into the LP letter or board deck — e.g., 'Management fees were $18k lower than March because two new fund closings were delayed to April'

Set up a Slack alert every 1st of the month that tells me QuickBooks and Plaid have finished syncing and links me to the flux dashboard so I know it's ready to review

Run these in Starch → or paste them into your favorite agent

Walkthrough

Step-by-step

1 Connect QuickBooks in Starch — Starch syncs your chart of accounts, invoices, bills, payments, vendors, and journal entries on a schedule. This becomes the actuals spine for every variance calculation.

2 Connect Plaid in Starch — Starch syncs categorized transactions and account balances from your operating and fund accounts daily, so cash movements are captured even when QuickBooks hasn't been fully reconciled yet.

3 Install the Runway Analysis starter app and tell Starch to extend it: 'Add a month-over-month P&L comparison view that shows actuals vs. prior month for every expense category, with variance in dollars and percent.'

4 Install the Budgeting starter app and load your annual budget by category — either by describing the allocations in natural language or by having Starch pull suggested allocations from your last 12 months of QuickBooks actuals.

5 Tell Starch: 'Build a variance flagging layer that highlights every row where actuals differ from budget by more than 10% or $5,000 in absolute dollars, and sorts the table by largest absolute variance first.'

6 Add a narrative generation step: 'For each flagged line item, draft a one-sentence explanation formatted for an LP update — keep it factual and reference the specific account and time period.'

7 Use Transaction Insights to cross-check Plaid-level spend against QuickBooks categories — if a vendor is appearing in Plaid but not reconciled in QuickBooks, the anomaly flag surfaces it before your CFO asks about it.

8 Set up a monthly automation: 'On the 2nd of each month at 9am, check that QuickBooks and Plaid have completed their latest sync, then send me a Slack message with a link to the flux dashboard and a summary of the top 3 largest variances.'

9 Build an LP-facing summary view: 'Create a clean one-page snapshot showing total revenue, total operating expenses, net burn, and the three largest favorable and unfavorable variances versus prior quarter — formatted for copy-paste into the quarterly letter.'

10 For portfolio companies that report through LP portals like Juniper Square or iLevel, tell Starch to automate pulling the monthly actuals through your browser and appending them to the flux dashboard — no API required.

11 Review the flagged variances with your GP or CFO using the narrative draft as a starting point — edit in Starch, and the updated language flows into the LP letter template automatically.

12 Archive each month's flux snapshot so you have a rolling 12-month history of explanations — useful when LPs ask in Q3 why Q1 management fee income was light.

See this running on Starch

Connect your tools, describe what you want, and the agent builds it. Closed beta is free.

Try it on Starch →

Worked example

March 2026 Monthly Close — $120M AUM Mid-Size Fund

Sample numbers from a real run

Management Fees Earned	214,000
Management Fees — Budget	250,000
Variance (Unfavorable)	-36,000
Operating Expenses — Actual	88,000
Operating Expenses — Prior Month	74,000
Variance (Unfavorable)	14,000
Net Burn — March	-127,000
Net Burn — February	-109,000

In March 2026, your flux dashboard flags two lines immediately. Management fees came in at $214k against a $250k budget — a $36k unfavorable variance. Starch's narrative draft reads: 'Management fees were $36k below March budget because the final close on Fund III was pushed to April 3rd, shifting $36k of fee recognition by one period.' That sentence goes straight into the GP update without editing. On the expense side, operating expenses hit $88k versus $74k in February — a $14k increase. Starch flags it and traces it to two new items in Plaid: a $9,200 charge from a data room provider and a $4,800 legal invoice, both one-time. The Transaction Insights anomaly detector caught the data room charge on March 14th because it was a vendor that had never appeared before — you saw it two weeks before close rather than discovering it mid-reconciliation. Net burn for the month is $127k versus $109k in February, and your Runway Analysis dashboard shows you're still at 26 months of runway at current pace. The entire flux review takes 40 minutes instead of a half-day.

Measurement

How you'll know it's working

Month-over-month variance by P&L line item, in dollars and percent, versus both prior month and annual budget

Management fee recognition timing — planned vs. actual by close date, flagged when a closing slip pushes fee income across a period boundary

Operating expense run rate versus prior quarter, broken down by category (legal, data room, travel, personnel)

Number of unreconciled Plaid transactions versus QuickBooks at time of close — a proxy for how clean the books actually are

Days to close — how many calendar days between month-end and a reviewed, LP-ready flux summary

Comparison

What this replaces

The other ways teams handle this today, and how the Starch version compares.

Excel + QuickBooks manual export

You control every cell, but every month-end starts with a fresh export, a broken formula, and a version-control problem when the CFO edits the shared file at the same time you do.

Juniper Square or Addepar

Purpose-built for fund IR reporting and deeply integrated with LP portals, but costs $50k+ per year, assumes a dedicated IR-ops headcount, and doesn't help you build the custom variance narratives or cross-system dashboards that live outside their template.

Google Sheets + connected QuickBooks add-on

Free and flexible, but data refresh is manual or semi-manual, anomaly detection doesn't exist, and you're still writing every variance sentence by hand.

Mosaic or Cube FP&A platforms

Strong for companies with a full FP&A function that wants to model scenarios and do headcount planning, but priced and scoped for teams larger than two and require implementation time your team doesn't have.

On Starch RECOMMENDED

One platform — runway analysis, quarterly budgeting, transaction insights all running on connected data. Setup in plain English; numbers stay current via scheduled syncs and live agent queries.

Try it on Starch →

FAQ

Frequently asked questions

Does Starch pull the full QuickBooks P&L or just transaction-level data?

Starch syncs QuickBooks entity-level data on a schedule — invoices, bills, payments, vendors, journal entries — and you build the P&L view on top of that. One honest note: QuickBooks report views (the pre-built P&L, Transaction List, and Vendor Expenses reports) are temporarily disabled pending an upstream connector fix. Entity-level data syncs normally, which means you can construct a flux analysis from the underlying records — it just means Starch is pulling the ingredients rather than the finished QuickBooks report.

We also use NetSuite for some portfolio companies. Can Starch reach that too?

Yes. Starch syncs your NetSuite data on a schedule — invoices, expenses, journal entries, balance sheets, and income statements. The Investor Reporting starter app uses NetSuite as a primary connection, so you can wire both QuickBooks and NetSuite into the same flux dashboard if your fund and portfolio companies run on different systems.

Our LP portal is Juniper Square. Can Starch pull data from it?

Juniper Square doesn't have a direct integration in Starch's catalog today. But Starch automates it through your browser — no API needed. You tell Starch which reports to pull, it navigates your Juniper Square account, extracts the data, and appends it to your flux dashboard. It works the same way for iLevel or any other portal you can log into.

Is Starch SOC 2 Type II certified? Our LPs will ask.

Not yet — Starch is not currently SOC 2 Type II certified. That's the honest answer. If your LPs or compliance team require SOC 2 Type II for any tool touching fund data, that's a real constraint to weigh.

Can Starch replace our fund admin's reporting?

No, and it's not trying to. Starch connects to your fund admin's outputs — QuickBooks, NetSuite, Plaid, and LP portals via browser automation — and lets you build the surfaces your team actually works in on top of that data. The fund admin closes the books; Starch makes the closed books usable for flux analysis, LP letters, and GP dashboards without a week of manual work.

What if I want to track variances against the budget I sent LPs three months ago, not just the current budget?

You can tell Starch to store the original budget snapshot and use it as the comparison baseline. Describe it: 'Compare March actuals against the Q1 budget we locked on January 15th, not the current budget version — and flag any line where we've revised the budget more than 15% since then.' Starch builds that view; you're not locked into comparing against whatever the current spreadsheet says.