Español

Financial Analyst Tools and Tech Stack: What FP&A Actually Runs On

Most FP&A stacks look the same from the inside. There's Excel. There's a folder of four spreadsheets someone pulls from NetSuite on the 5th of the month. There's a BI dashboard nobody trusts because last quarter's revenue line moved when the data team "fixed" a join. The board deck takes nine days. The CFO asks one new question on Thursday and the analyst loses the weekend.

I've seen this stack at Series B startups and at $400M business units inside Fortune 500s. The tools change. The pain doesn't. And here's the thing most vendors won't tell you: this isn't a tooling problem yet. It's a clarity problem about what each tool category should actually do.

So before you sit through another Workday Adaptive demo, let's go category by category. Real prices. Real names. What it's for, what it's not for, and when you actually need to upgrade.

The Core 6 — what every FP&A stack actually needs

There are six jobs the stack has to do. Not six tools necessarily, but six jobs. Modeling, planning, raw data access, visualization, the close, and revenue data. Skip one and something breaks. Double up and you're paying twice for the same outcome.

1. Spreadsheet and modeling layer

Excel plus Google Sheets. Don't apologize for it.

Excel with proper macros, Power Query, and a disciplined naming convention beats most $50K/yr planning tools for any company under 200 headcount. I'm not being contrarian. The reason is simple: planning tools force you to model their way. Excel lets you model your way. For an analyst who actually understands the business, that flexibility is worth more than any "single source of truth" pitch.

Use Sheets for live collaboration, the IMPORTRANGE pulls, anything a board member or a non-finance partner needs to touch. Use Excel for the heavy modeling — pivot tables on 200K-row exports, scenario tabs, the actual three-statement model. The modern Excel with LET, LAMBDA, and dynamic arrays is a different product than the one most analysts learned in 2018. If your team is still building VLOOKUP chains, that's a training problem, not a tooling problem.

Camellia would say: spend $2,000 on a senior analyst's Power Query training before you spend $80K on a planning tool. You'll get more lift.

2. Planning and xP&A tool

This is where most teams overspend. The category is real, but the price ranges are wide and the right answer is rarely the most expensive option.

  • Workday Adaptive Planning: enterprise-grade, often $80K to $200K per year by seat count and module. Worth it if you're publicly listed, or you have 500+ headcount across multiple entities, or your audit committee explicitly asked for an auditable planning system. Otherwise it's overkill.
  • Mosaic: roughly $30K to $100K per year. Strong on SaaS metrics, opinionated UI, fast to deploy. If you're a SaaS company between $20M and $100M in ARR and you actually want to leave Excel, this is usually the most defensible choice.
  • Cube: entry pricing around $1,500 a month and up. Excel-native, which means your team keeps modeling in spreadsheets and Cube acts as the planning layer underneath. The right pick for teams that refuse to leave Excel and shouldn't have to.
  • Pigment: newer, multi-dimensional, real deployments land around $50K and up. Very flexible, sometimes too flexible. Pick this if your business has unusual dimensionality (multi-currency, multi-entity, complex allocations) and you have a finance systems person to actually configure it.

The honest filter: which one will your team actually use? Most planning tools fail not because the tool is bad, but because the analyst hates leaving Excel and the implementation never finishes. Pick based on adoption probability, not feature depth.

3. Data warehouse access

Snowflake or BigQuery. Sometimes Redshift, less often these days.

The job here isn't for the FA to be a data engineer. The job is to have read access to a warehouse, with a few governed views, so the analyst can pull what they need without filing a ticket and waiting three days for a CSV. If your analyst is still emailing the data team for monthly cohort exports, your stack has a problem the planning tool can't fix.

Cost is downstream of usage. For a small org running a few dashboards and ad-hoc queries, budget $1K to $5K a month. The bigger cost is human: someone has to build and maintain the views. That's usually a shared analytics engineer, not a finance hire.

Camellia would say: if you can't pull last quarter's revenue by segment in under 60 seconds without bothering anyone, fix the data access before you buy anything else.

4. BI layer

Pick one. Don't run all three.

  • Looker (Google): expensive, governance-heavy, great when finance and the rest of the company need to share definitions. Pricing isn't public for a reason. Budget enterprise.
  • Tableau: about $70 per user per month at the Creator tier. Strongest visualization, weakest governance. The right pick if your CFO presents to the board with charts.
  • Hex: about $50 per user per month. Notebook-first, fast for ad-hoc analysis, plays well with Snowflake and BigQuery. The right pick if your analysts want to write SQL and Python in the same place a chart lives.

The pattern I see most often: a company buys Looker for governance, then the analysts buy Hex with their own corporate cards because Looker is slow for exploratory work, and now the company is paying for both. That's not a stack, that's a tax.

Pick one for the company. If analysts need a notebook tool, pick Hex and skip Tableau for finance entirely. The viz gap is smaller than you think.

5. Close and actuals system

NetSuite, Sage Intacct, or QuickBooks at smaller scale. Whatever your controller picked, you're stuck with it. Don't get into that argument as the analyst.

What matters for the FA is: clean API or scheduled export. If pulling actuals means a controller hits "export to Excel" and emails a trial balance to you on the 7th, your stack is broken at the foundation. The fix isn't a planning tool. The fix is automating the data pull, usually a Fivetran connector or a custom script into the warehouse. Once actuals land in the warehouse on a schedule, every downstream report becomes faster and more honest.

The FA's job here is the data pull, not the close itself. Stay in your lane on this one. The controller owns the close.

6. CRM for revenue data

This is the line where most companies overpay and the analyst has the least leverage.

Pipeline, ARR, sales forecast, won/lost reasons: they all live in the CRM. The FA needs read access and a stable schema. They do not need a seat that lets them edit deals. That's a sales seat. Different job.

The pricing reality:

  • Rework CRM + Sales Ops: starts at $12 per user per month. Pricing at rework.com/pricing. For a finance team that needs read access to pipeline and revenue, this is the cheapest defensible option that still has a real schema and an API.
  • HubSpot Sales Hub: $90+ per user per month at the Professional tier most analysts actually need for reporting depth.
  • Salesforce: $165+ per user per month at the Enterprise tier, before any of the add-ons that make the reporting actually work.

Now, full disclosure: I work for Rework, so take the recommendation with the appropriate salt. But the math holds either way. If your sales org is on Salesforce and your finance team needs read access for forecasting, that's not a question. You stay on Salesforce because the data is there. The question is what you put on top for the workflow your sales team actually runs day to day. And that's where the $12 vs $165 gap matters.

For a small or mid-size org standardizing fresh, Rework lands at $12 because we don't tier reporting behind a Pro upgrade. For an established Salesforce shop, the right answer is usually a warehouse connector that pulls Salesforce objects nightly into Snowflake, and you skip giving finance a Salesforce seat at all.

Either way: don't pay enterprise CRM rates to extract revenue data. That's what cheaper CRMs and warehouse connectors are for.

The 30-day FP&A stack audit

Run this before you buy anything. If you can't write the recommendation page at the end, you don't know your own stack well enough to upgrade it.

Days 1–5: Inventory. List every spreadsheet, every dashboard, every recurring export, every scheduled report. For each one: who owns it, who consumes it, how often it breaks. You will find at least three things nobody actually reads anymore. Kill them in week four.

Days 6–10: Map data lineage. For every number on the board deck, trace it back to the source system. Most companies have two or three numbers that nobody can fully trace, usually a "growth rate" or a "margin" calculated three layers deep in someone's tab. Write down each one. You'll fix half of them just by mapping them.

Days 11–15: Cost the stack. Pull every finance-tool invoice for the last 12 months. Calculate cost per analyst, cost per scenario run, cost per close. Compare against headcount growth. If the stack cost grew faster than the team did, something is wrong.

Days 16–20: Identify the top 3 "manual misery" tasks. The ones the analyst dreads on a recurring basis. Time them honestly. For each: is this a tooling problem (the right software would eliminate it), a process problem (we're doing the wrong steps), or a data-access problem (we're rebuilding what the warehouse already has)? Be honest. Most "tooling" problems are actually process problems wearing a tooling costume.

Days 21–25: Pilot one fix. Just one. A Power Query refresh that replaces a Friday spreadsheet. A Hex notebook for the variance commentary. A scheduled warehouse view for the cohort report. Pick the smallest fix with the highest weekly time savings and ship it. Nothing else.

Days 26–30: Write a one-page recommendation. What to keep. What to consolidate. What to evaluate next quarter. What to kill outright. Send it to the CFO before they ask. This page is the artifact that turns you from "the analyst" into "the analyst who runs FP&A." Don't skip it.

When to graduate from Excel

This is the section everyone wants. Three thresholds. Cross any one of them and the conversation changes.

  • Threshold 1, 200+ scenario versions per year. Board cases, reforecasts, what-ifs, sensitivities. At this volume, Excel becomes unauditable. You can't tell which case the CFO actually approved.
  • Threshold 2, 100+ cost centers or 50+ entities. Manual roll-ups break. Currency translation breaks. Inter-company eliminations break. You need a tool with a real dimensional model.
  • Threshold 3, three or more people edit the model every week. Version conflicts cost more than the tool does. Someone overwrites the rev forecast. Someone forgets to copy down the formula. The "shared spreadsheet" tax is real and it's bigger than $50K a year past this point.

Below these thresholds: stay on Excel. Fix the macros. Hire one person who actually knows Power Query and dynamic arrays. You'll save $80K a year and your model will run faster than the planning tool would have.

What to skip — Camellia's hot takes

A few rules I've watched teams break and regret:

  • Don't buy a planning tool to fix a process problem. If the model is wrong on Excel, it'll be wrong on Adaptive. The planning tool will just give the wrong answer faster, with a prettier UI, and on a renewable contract.
  • Don't centralize on a BI tool the analyst hates. Adoption is everything. A "governed" Looker instance that nobody opens is worth less than a chaotic Hex deployment everyone uses.
  • Don't pay enterprise CRM rates just to extract revenue data. A warehouse connector and a $12 CRM beat a $165 CRM you only use for reports.
  • Don't run a "stack consolidation" project in the same quarter you're closing books on a new entity, integrating a new acquisition, or onboarding a new CFO. Pick your battles. FP&A is a marathon and the stack will still be there next quarter.

Closing

The FP&A stack isn't six tools. It's six jobs that need to get done — modeling, planning, raw data, visualization, actuals, and revenue. Most teams have the tools and lack the wiring. The board deck takes nine days not because the planning tool is bad, but because the data doesn't flow on its own and three of the six jobs are stitched together by hand on the morning of the 5th.

Audit first. Wire second. Buy third. In that order, and not the other way around.

Learn More