Replacing an Excel Quote Calculator: CPQ Software or a Custom Tool?

TL;DR

• Separate the calculation from the surrounding work: choosing valid products, applying customer terms, approving exceptions and recording the offer sent to the buyer.
• Check discount bases, rule precedence and quantity thresholds explicitly. Two systems can display the same percentage and calculate different prices.
• Test existing CRM quoting, specialist CPQ and a narrower calculation extension before commissioning a complete replacement.
• Keep an issued quote tied to its own inputs, prices and revision. Updating the price list should not silently rewrite an earlier offer.
• Give candidates the same five representative quotes, then ask the future administrator to make a price change. Include configuration and ongoing rule maintenance in the comparison.

What does your quote calculator actually do?

For example, suppose a sign-making business quotes panels with different dimensions and finishes. Sales enters the specification in Excel, chooses customer terms and sends unusual discounts to the director. Someone then copies the approved amount into a proposal and the accepted specification into production.

Replacing that workbook involves more than reproducing its total. The business needs to know whether the selected combination is sellable, which prices apply, who may override them and which version the customer received. This scenario and the calculations below are hypothetical, not a Caemcore case study.

CPQ means configure, price, quote. HubSpot's CPQ documentation, for example, covers product configuration, quote creation and approvals within the CRM. A ready-made product may already handle much of the surrounding workflow. A custom formula is not evidence that all of it must be rebuilt.

Separate the workbook's responsibilities before choosing a replacement. Product compatibility, estimated production cost, selling price and permission to offer a discount are different rules. A model can calculate the right amount while leaving the wrong person free to change it.

Establish the rules behind a correct quote

Preserve an unchanged copy of the workbook, the price data it uses and representative offers already sent. Work on copies during the assessment. Ask a salesperson and the pricing owner to explain an ordinary quote and a negotiated exception, including any values they enter over a formula.

Record the calculation settings too. Microsoft documents automatic and manual recalculation in Excel. A saved displayed value should not automatically become the expected result for a replacement. Recalculate the test copy with the agreed inputs, and have the pricing owner resolve discrepancies between the workbook, the intended rule and the offer actually sent.

For each important rule, write down its inputs and units, the applicable product or customer, its effective date, its position in the calculation and the person allowed to change it. Specify what happens when an input is missing. An unknown delivery charge should not quietly become zero because the new form requires a number.

Keep cost and selling price distinct. A production-cost estimate can inform a margin check without becoming the price offered to the customer. The rule must say which costs and charges enter that check; a field named margin does not settle the definition.

Compare four routes, not just Excel versus custom code

Use the same rules and quote examples for every route. The following table identifies what would make each option worth investigating, rather than declaring one suitable from the size of the workbook.

A retained workbook can be a reasonable answer when the process is small and controlled. Replacing it is worth investigating when essential checks depend on someone remembering an exception, the team cannot identify the current offer, or maintaining the surrounding work has become unacceptable. Measure the actual problem; do not infer it from the file extension.

Check the CPQ already available to you

Start with the CRM or business platform you use, then consider specialist CPQ products when their model better matches the work. HubSpot's current CPQ overview lists Revenue Hub Professional and Enterprise, with seats required for certain functions. Include the actual subscription and users in a proposal, not just a feature shown in a vendor's account.

Its product-pricing documentation distinguishes volume pricing, where the selected unit price applies to all units, from graduated pricing, where each band prices only the units within it. Both can be called quantity pricing in conversation. Have the supplier demonstrate which rule it has configured.

Zoho CRM's CPQ documentation describes product configuration and conditional pricing, including direct and volume-based prices or discounts. It also states that when several price rules match a product, only the first matching rule in sequence applies. That matters when your workbook combines customer and quantity discounts: two matching rules do not automatically mean two discounts will be applied. Request a demonstrated configuration for the intended result, rather than concluding that the product is either suitable or unsuitable from the feature name.

Distinguish a warning, a prohibited offer and an exception that needs approval. HubSpot's quote rules, documented for Revenue Hub Enterprise, can warn or block quote publication. Its approval setup separately controls which quotes require review and by whom. A warning the salesperson can ignore does not enforce a rule that the business says must stop the offer.

These are documentation examples, not a product ranking or hands-on comparison. Check feature availability and the configured behavior in the account you would use.

Keep the calculation model when that is the valuable part

A workbook-backed application can separate a spreadsheet model from the interface employees use. SpreadsheetWeb, for example, describes running uploaded workbook formulas through its calculation engine, with web inputs connected to cells. Its documentation states that VBA does not execute. A workbook containing macros therefore needs a specific assessment of those steps, not a promise that uploading it reproduces the whole process.

Test the functions, external data and recalculation behavior your model actually needs. Also test permissions, quote storage and what happens when the model is updated. Preserving formulas does not automatically preserve the surrounding approval history or make the application a complete CPQ system.

Worked example: the discount base changes the answer

Hypothetical calculation. All amounts and commercial rules in this example are invented to demonstrate a test. They are not recommended prices, vendor fees or client results.

Suppose the quote covers ten panels, each with a finished area of 2 m². The base selling rate is $90 per m², finish costs the buyer another $20 per panel, and production preparation is a $150 charge per quote. Tax and delivery are excluded from this illustration.

The business's hypothetical policy gives a 5% quantity discount on the base panels and finish when at least ten panels are ordered. An eligible customer's 10% discount then applies to that discounted subtotal. Neither discount applies to preparation. For this example, round each quote-level discount to cents and keep the intermediate amounts.

The total is (10 × 2 × $90 + 10 × $20) × 0.95 × 0.90 + $150 = $1,860.

Applying a single 15% reduction instead gives $1,850. Applying both discounts to the preparation charge as well gives $1,838.25. Either implementation could display “5% volume, 10% customer” and still be wrong for the agreed policy.

Test the quantity boundary too. With all other assumptions unchanged, nine panels total $1,770, ten total $1,860, and eleven total $2,031. Confirm whether the discount applies to every unit once the threshold is reached or only to units above it. Test whether separate lines for the same eligible product count together or separately.

The acceptance evidence is the calculation trace: inputs, eligible lines, rule sequence, intermediate amounts and final total. A supplier manually entering $1,860 has demonstrated data entry, not the pricing rule.

Give every candidate the same five representative quotes

Choose examples from your actual work and remove customer-identifying information before sharing them. Prepare the expected results with the pricing owner. Where a needed exception has no historical example, create a clearly labelled test case. Run the exercise in a non-production environment without sending offers to customers.

1. The ordinary quote. Use a normal specification, customer and current price list. Have the salesperson create the offer from the original enquiry. Check quantities, units, charges and the customer-facing document against the approved calculation. Record any manual work still required outside the tool.
2. An invalid or incomplete configuration. Try a finish that your catalogue excludes for the selected product, then omit a required dimension. Check whether the result is a refusal, a request for information or an authorized review. An incomplete specification must not produce a normal-looking final price through an unexplained default.
3. A threshold and a customer-specific rule. Use a quote near a quantity break, with negotiated customer terms. Test just below, at and above the boundary. Inspect which rules combine, which take precedence and which charges are excluded. The worked example above shows the detail to request.
4. A discretionary discount. Start from the calculated price and ask for an additional concession that requires approval under your policy. Confirm who can approve it, what they see and whether the salesperson can issue it beforehand. Change the price or specification after approval and test the agreed rule for renewed review. Keep the exception reason and approved revision.
5. A revised offer after a price-list change. Issue a test quote, update the underlying price list, then open the original and create a revision. Verify which values remain fixed and which recalculate. Accept the intended version through the test process and check that the saved order specification and amount match it, rather than a newer draft or today's catalogue.

Record passed, failed and untested results with the configuration used. A failed demonstration may identify a missing setting, an unresolved business rule or a genuine product limit. Ask the implementer to distinguish those causes before recommending custom development.

For access controls, include a technical check of the protected action. OWASP recommends validating authorization on every request. A hidden discount field is not sufficient evidence that an unauthorized price change will be refused. These workflow checks do not replace a complete security assessment.

Keep the issued quote separate from the latest model

Specify the record you need to retrieve later: quote identifier and revision, customer, configuration, quantities, applicable price-list or model version, calculated amounts, authorized overrides, approval and the terms sent to the buyer. Preserve the actual customer-facing document or equivalent immutable representation too.

Separate calculated, internally approved, issued and accepted states. The business may allow an unissued draft to recalculate when a price changes, while requiring an issued offer to retain its terms until revised or withdrawn. Write that policy explicitly. An approval of version one should not silently authorize a materially different version two.

The order handoff should identify the accepted quote revision and its specification, not just copy the latest total from the CRM. On a failed transfer, retain the accepted record and make the pending work visible. That connection needs its own acceptance check even when both applications calculate the same amount.

When custom development is justified

Custom work deserves a proposal when a required rule or workflow remains unsupported after a competent configuration check, or when the supported workaround creates an operating burden the business will not accept. State the gap concretely: for example, the approved model needs a production calculation and traceable inputs that the evaluated product cannot represent without repeated manual intervention.

Ask whether that gap needs a calculation component rather than a complete quoting application. A custom service could return the configuration, calculated lines and model version while a suitable CPQ retains document generation and approvals. Verify that the receiving product supports the required fields and operations before choosing that split.

A full custom tool needs a broader reason, such as an essential estimating workflow, user interface and data model that cannot reasonably fit the available products. More products, discount rules or spreadsheet tabs do not establish that reason by themselves. Equally, a promise that a product supports unlimited rules does not show whether your staff can maintain the proposed configuration.

Include an administrative test: have the intended owner change a rate for a future date, preview its effects on known quotes and put the approved change into use. Establish who may make that change and how an incorrect new rate is corrected without rewriting issued offers. This tests whether the solution will remain usable after the original implementer leaves.

Compare the whole commitment, then pilot one product family

Request an implementation scope that includes extracting and agreeing the rules, product and customer data, configuration or development, quote documents, approvals, integration, testing and handover. Separate ongoing subscriptions, hosting and support. Include the time your own pricing owner will spend checking and maintaining the rules.

Compare the same release and operating period. A lower subscription can still require substantial specialist configuration; custom code still needs maintenance. Our custom versus off-the-shelf guide explains how to compare cash spending and employee time without counting released hours as automatic payroll savings.

Pilot one representative product family and its meaningful exceptions. Compare outputs before allowing the new tool to send real offers, and designate one authoritative sending route during the transition. Keep a controlled fallback so a failed calculation does not force staff to improvise a customer price.

For the broader data move, our spreadsheet migration guide covers choosing the records and workflow that need to move first. Do not expand the quote-calculator project into replacing every workbook in the company.

Choose CPQ when the tested configuration and its operating cost fit. Retain the workbook model or build a narrower component when that closes the gap. Commission a full custom quoting tool when you can name what the smaller options leave unresolved and show how the proposed release will handle it.

About Caemcore

Caemcore builds custom software for companies that outgrew off-the-shelf tools.

Internal systems, integrations, web and mobile apps, and products built from scratch. Small senior team, Warsaw-based, working with clients across the EU, US and UK.

You get a price range and a timeline before development starts, a working demo every Thursday, and a final invoice that matches the number quoted on day one.

Not sure what you need? A 30-minute call, free: what the task actually is, whether it needs building or an off-the-shelf tool covers it, a range and a timeline. You get that whether we work together or not.

FAQ

What if nobody can explain an important formula in the old workbook?

Treat the intended rule as unresolved. Preserve the formula, find examples of its use and ask the person responsible for pricing to approve the expected result. A developer can explain what the formula calculates, but cannot establish commercial intent from the code alone. Record any deliberate correction separately from a change in how the calculation is implemented.

Can customers use the same calculator as the sales team?

That is a separate access and product requirement. Decide whether customers receive a budget estimate, submit a request for review or obtain an offer they can accept. Restrict the fields and results they can access, including internal costs and negotiated prices. Test the customer journey independently; hiding internal controls in a public page does not establish that the underlying information is protected.

How should the system handle a completely bespoke request?

Provide a controlled route to a named estimator rather than inventing a price from the nearest catalogue item. Record the missing specification, the estimator's assumptions and the approval required to issue the result. Label that work as manually estimated. Review recurring exceptions later to decide which deserve a reusable rule, without forcing every unusual job through an automatic calculation.

Do all historical quotes need to move into the new system?

Separate active offers from closed history. Active offers need their original terms, revision and status, plus a workable route for later acceptance or amendment. Closed quotes may remain in an accessible archive if that meets the business's needs and retention requirements. Do not reprice historical documents with today's rules merely to make them fit the new model.

What changes when we quote in more than one currency?

Specify whether each currency has its own price list or uses conversion, then identify the rate source, effective date and rounding rule. Preserve the currency and rate used on an issued quote. Decide separately how a renewal or revised offer is priced. Include tax handling and the currency accepted by the order system in the scope, with the responsible finance team approving those rules.

Sources

• HubSpot, Understand HubSpot CPQ. Checked 15 September 2026.
• Microsoft Learn, Excel performance: Improving calculation performance. Checked 15 September 2026.
• HubSpot, Create and manage products. Checked 15 September 2026.
• Zoho CRM, CPQ: How it works. Checked 15 September 2026.
• HubSpot, Configure and manage quote rules. Checked 15 September 2026.
• HubSpot, Set up quote approvals. Checked 15 September 2026.
• SpreadsheetWeb, Excel models as web applications. Checked 15 September 2026.
• OWASP, Authorization Cheat Sheet. Checked 15 September 2026.

Product capabilities were checked on 15 September 2026 using vendor documentation, not hands-on comparison tests. The sign-making business, prices, discount policy and test scenarios are hypothetical editorial examples, not supplier quotations, recommended pricing or Caemcore client results. The selection criteria are decision aids, not universal thresholds for custom development.