How to run monthly flux and variance analysis as Chief of Staff and Founder's Office

Finance & FP&AFor Chief of Staff and Founder's Office3 apps12 steps~24 min to set up

You're the one who has to explain why engineering payroll was 22% over budget in March before the CEO walks into the board call at 2pm. You pull the QuickBooks report, cross-reference it against the Plaid transactions, open three different Notion docs the functional leads sent you, and try to reconcile why the numbers don't match. The BI tool the last analyst set up hasn't been touched in four months and nobody knows the login. Monthly flux analysis — comparing actuals to prior month and to budget, then writing the narrative — takes you the better part of a day, mostly in spreadsheets, and the output is a Google Sheet that's stale by the time anyone reads it.

Finance & FP&AFor Chief of Staff and Founder's Office3 apps12 steps~24 min to set up
Outcome

What you'll set up

A live variance dashboard that pulls actuals from QuickBooks and Plaid on a schedule, compares them to your quarterly budget by category, and flags every line item that moved more than 10% month-over-month — no manual exports required.
An automated monthly narrative generator that reads the variance data and drafts a plain-English explanation of the top movers — ready to drop into the board deck or investor update in minutes, not hours.
A recurring Slack alert every first business day of the month that surfaces the three largest budget variances so you and the CEO see the story before anyone asks for it.
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.

Data sources & config

Starch syncs your QuickBooks data on a schedule — invoices, bills, payments, vendors, and journal entries refresh automatically, giving the variance dashboard always-current actuals without manual exports. Starch also syncs your Plaid bank feed on a schedule for transaction-level detail and cash reconciliation. Slack is connected from Starch's integration catalog and the agent queries it live to post the monthly narrative. Your budget targets live inside the Budgeting app and are compared against QuickBooks actuals automatically each cycle.

Prompts to copy
Build me a monthly flux analysis dashboard that compares QuickBooks actuals by category to my quarterly budget targets. Show me dollar variance and percent variance for each line, flag anything over 10% in red, and include a prior-month comparison column so I can see the trend.
Every first business day of the month, generate a plain-English narrative summarizing the three biggest budget variances from last month — what moved, by how much, and what the likely cause is based on the transaction detail. Post it to the #exec-team Slack channel.
Show me a month-over-month spend trend by vendor for the top 20 vendors by spend, pulling from Plaid transactions. Highlight any vendor where this month's charges are more than 15% above the 3-month average.
Run these in Starch → or paste them into your favorite agent
Walkthrough

Step-by-step

1 Connect QuickBooks in Starch — Starch syncs your invoices, bills, vendor payments, and journal entries on a schedule. No exports, no CSV uploads.
2 Connect Plaid so Starch syncs your bank transactions on a schedule alongside the QuickBooks data. This gives you both the accounting view and the cash view in one place.
3 Open the Budgeting starter app and enter your quarterly budget targets by category — headcount, software, contractors, marketing, office, travel. If you've already got a budget in a spreadsheet, paste the category totals in.
4 Tell Starch: 'Build me a monthly flux table that shows actuals vs. budget and actuals vs. prior month for every QuickBooks expense category, with dollar variance and percent variance columns. Flag any line over 10% variance in either direction.' Starch assembles the app.
5 Open the Runway Analysis starter app and wire it to the same Plaid and QuickBooks connections. This gives you the cash-burn context alongside the variance data — so when engineering payroll is up $40k, you immediately see what that does to your runway number.
6 Use Transaction Insights to drill into the specific vendor-level detail behind any flagged category. If 'Software & Subscriptions' is 18% over, Transaction Insights shows you which three vendors drove it and whether those charges are new or recurring.
7 Connect Slack from Starch's integration catalog, then tell Starch: 'Every first business day of the month, read last month's variance data, identify the three largest movers, write a two-paragraph narrative explaining what changed and why, and post it to #exec-team.' Starch schedules and runs this automatically.
8 Build a secondary view scoped to headcount cost: 'Show me total compensation spend from QuickBooks bills and payroll entries, broken out by department, compared to prior month and to the quarterly headcount budget. Flag any department over 5% variance.' This is the view you share with functional leads for accountability.
9 Before the monthly board prep call, open the flux dashboard and copy the narrative Starch generated. Paste it into the investor update or board deck as the first draft of the financial commentary — edit for tone, not for facts.
10 If a variance needs a follow-up, tell Starch: 'Find all QuickBooks bills from [vendor name] in the last 90 days and show me the amounts and dates.' The agent queries the synced data live and surfaces the detail in seconds.
11 Set a second automation: 'If any single QuickBooks expense category comes in more than 20% over budget mid-month, send me a Slack DM immediately — don't wait for the monthly run.' This catches surprises before close, not after.
12 Share the dashboard link with the CFO or finance lead so they're looking at the same live data you are — no more 'which version of the spreadsheet is current' conversations before board calls.

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 — Series B Company, 150 Employees

Sample numbers from a real run
Engineering Payroll & Contractors487,000
Engineering Payroll Budget410,000
Variance77,000
Software & Subscriptions (Actual)62,400
Software & Subscriptions (Budget)48,000
Variance14,400
Sales & Marketing (Actual)193,000
Sales & Marketing (Budget)210,000
Variance-17,000
Travel & Offsite (Actual)38,200
Travel & Offsite (Budget)22,000
Variance16,200

The March flux dashboard flags four lines immediately. Engineering payroll came in $77k over budget — the largest single variance. Starch pulls the QuickBooks bill detail and surfaces the cause: two contract developers onboarded mid-February whose first full month hit in March, plus a contractor extension that wasn't reflected in the original quarterly budget. You add a note, fix the forward budget, and the runway model updates automatically — March's overrun compresses the 18-month projection to 16.4 months, which the CEO needs to know before talking to investors. Software and subscriptions are $14.4k over; Transaction Insights shows a new $8,200/month data vendor the head of engineering approved and two seat expansions on tools that crossed a pricing tier. Travel is $16.2k over budget because the exec offsite landed in March instead of April as originally planned — a timing issue, not a structural overage, and the Starch narrative flags it as such. Sales and marketing came in $17k under, which looks good until you check the pipeline data and realize two planned campaigns got pushed. The narrative Starch drafted goes into the board deck as-is, with one sentence edited by you for tone. Total time from 'QuickBooks sync complete' to 'narrative in the deck': 25 minutes.

Measurement

How you'll know it's working

Monthly budget variance by category ($ and %) — the primary table the CEO and board review
Month-over-month expense change by department — the accountability metric you share with functional leads
Runway impact of actuals vs. budget — how much the current month's variances shift the forward cash projection
New vendor spend detected in the period — catches unbudgeted tools and one-off charges before they become patterns
Time from books close to narrative delivered — the operational metric for whether this process is actually working
Comparison

What this replaces

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

QuickBooks reports + Google Sheets
You can export QuickBooks P&L reports and build a variance model in Sheets, but every month is a manual export-and-paste cycle, and the spreadsheet is stale the moment you close it — Starch keeps the data live and builds the narrative automatically.
Mosaic, Jirav, or Cube (FP&A tools)
Purpose-built FP&A platforms are powerful but assume you have a finance team to configure and maintain them; at 150 people with a chief of staff doing this job, the setup cost and seat pricing rarely pencil out compared to describing what you need and having Starch build it.
Looker or Mode (BI tools)
A BI tool can produce beautiful variance dashboards once a data engineer builds the models, but if the analyst who set it up has left, you're looking at a black box — Starch lets you describe the logic in plain English and own the output yourself.
Excel + Accountant-delivered close package
Waiting for your bookkeeper to deliver the close package means the flux analysis lands 10-15 days after month end, which is too late to catch surprises before they become board-level conversations.
On Starch RECOMMENDED

One platform — quarterly budgeting, runway analysis, 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 sync all of my QuickBooks data or just some of it?
Starch syncs 20+ QuickBooks entities on a schedule — invoices, bills, payments, vendors, journal entries, and more, up to 50,000 records per entity. One current limitation worth knowing: QuickBooks report views like the pre-built P&L and Transaction List are temporarily unavailable while an upstream connector issue is being fixed. The variance analysis in Starch is built from entity-level data directly, which is actually more flexible than the canned reports — you describe the grouping and comparison logic you want, and Starch assembles it.
What if my budget lives in a Google Sheet or Notion doc, not in a budgeting tool?
You can connect Google Sheets from Starch's integration catalog — the agent queries it live when your variance app runs, so your budget targets don't have to move. Alternatively, enter your category targets directly in the Budgeting app and Starch tracks the comparison from there. If your budget is in a Notion database, Starch syncs Notion on a schedule and can read the targets from there as well.
Can Starch actually write the board narrative, or does it just show me the numbers?
Both. The variance dashboard gives you the numbers in a clean table. The narrative automation — which you set up by describing what you want in plain English — reads the variance data and writes a plain-English explanation of the top movers each month, then posts it to Slack or saves it to Notion. You edit for tone; you don't have to draft from scratch. The output quality depends on how specific you are in the prompt — the more context you give about what 'normal' looks like for your business, the better the explanation.
Is my financial data secure? Is Starch SOC 2 certified?
Starch is not SOC 2 Type II certified yet — that's an honest answer and worth knowing if your company has compliance requirements that mandate it. There's no on-premises or self-hosted deployment option. If SOC 2 Type II is a hard requirement for your finance data today, that's a real constraint to weigh.
What if I need department-level variance, not just company-wide?
Tell Starch exactly that: 'Show me budget vs. actuals broken out by department, using the class or location tags in QuickBooks.' If your QuickBooks data has department coding on bills and expenses, Starch reads those fields and can slice the variance table however you define it. If the coding is inconsistent, Starch can flag uncoded transactions so you can see the gap.
We use NetSuite, not QuickBooks. Does this still work?
Yes. Starch syncs NetSuite data on a schedule — invoices, expenses, journal entries, balance sheets, and income statements. The flux analysis setup is the same; you're just pointing it at the NetSuite connection instead of QuickBooks.

Ready to run run monthly flux and variance analysis on Starch?

Request closed-beta access. Everything is free during beta.

You're on the list! We'll be in touch soon.