LightSource vs. Excel and email

Where Excel and email are genuinely the right tool

Nobody chose Excel and email for sourcing; the stack was already there. Roughly 750 million people use Excel, per Microsoft CEO Satya Nadella in 2017, and third-party estimates run above a billion. Every PLM and ERP exports to it, every cost engineer models in it, and every supplier on earth can open a bid sheet with no account and no training. Trade-press surveys put spreadsheet reliance among supply chain professionals at 65 to 80%, and two-thirds of companies say they consider Excel a supply chain system.

The strengths are real. A blank grid offers total modeling freedom: any bid sheet, any cost breakdown, any should-cost structure, with no vendor roadmap in the way. The marginal software cost is near zero, since Excel ships inside the Microsoft 365 subscription most manufacturers already own. Implementation is zero; it is already on every laptop, and it works offline in a plant with no Wi-Fi. The grid also keeps improving: Copilot Agent Mode, generally available on Excel for the web in December 2025 and on Windows desktop in January 2026, builds formulas, tables, and PivotTables from natural language, though it requires a paid Copilot license and it operates on cells rather than on parts, quotes, or suppliers.

When The Wall Street Journal reported in 2017 that finance chiefs were telling their teams to stop using Excel, the backlash was loud enough to earn its own follow-up coverage. That attachment is earned. For a team sourcing a small number of parts a year, Excel and email is arguably the correct tool.

Where LightSource wins

The wins are structural, and they compound with part count, supplier count, and time.

Parts instead of rows. LightSource keeps a revision-managed item master: nested BOMs, snapshots, diffs, and cost roll-ups. AI ingestion builds it from the Excel files, PDFs, and drawings a team already has, so parts stop being rows retyped into each new bid sheet.

RFQs that carry revision integrity. Drawings attach to the part, not to a ZIP on an email, so suppliers quote against the current revision and the record shows who quoted what, on which revision, and when.

Flexible cost breakdowns without hand-merging. In our evaluations, LightSource is the only sourcing solution with true flexible cost breakdowns: category-specific bid sheets per event, with a drag-and-drop editor and a formula editor that keep the modeling freedom buyers built in Excel. AI then normalizes every returned quote into one comparable structure, so nobody hand-merges eight differently mangled attachments.

Price history that belongs to the company. Awards become tracked contracts with per-part, per-supplier price history and cost walks. US median employee tenure is 3.9 years per the Bureau of Labor Statistics; when a buyer leaves, the record stays.

A network instead of an address book. 20,000 suppliers are live on the network, with AI capability-based discovery, and suppliers use it free.

A small implementation gap. The status quo's one structural advantage is zero rollout, and against LightSource that advantage is narrower than against any suite: live in days on flat files, sourcing in about 30 days, implementation included. Manufacturers including BRP, Canada Goose, HelloFresh, and Shure run direct materials sourcing on LightSource today; Gartner named it a Cool Vendor in 2025, and it carries venture backing co-led by Lightspeed and Bain Capital Ventures ($33M).

The structural case: why files, inboxes, and memory fail at scale

The status quo has four components: the grid, the inbox, the shared drive, and the buyer's memory. The fourth is the actual system of record for supplier history and past pricing, and it is the component that walks out the door.

Spreadsheet error rates are a documented research field. Raymond Panko of the University of Hawaii, surveying 85 intensive inspection studies of operational spreadsheets, found that 94% contained errors, with cell error rates of 1 to 5% on nontrivial work and average error-detection rates in inspection experiments of only about 60%. His framing is careful: "spreadsheet programs are not error-prone. People are error prone." Panko's point is about process, not software: a chain of hand-copied files gives human error nowhere to get caught, which is exactly why shared records and audit trails exist.

Most famous spreadsheet failures come from finance, academia, and public health rather than sourcing departments, and they should be read as general spreadsheet-risk evidence. One, though, is squarely a sourcing story: in 2003, TransAlta lost $24 million, about 10% of annual profit, when a cut-and-paste misalignment in a bid-submission spreadsheet attached bids to the wrong contracts. And Public Health England's 2020 loss of 15,841 COVID case records, silently truncated by a legacy .xls row limit, shows the failure mode at scale: the pipeline produced no error message, just missing rows.

Email adds its own measured cost. McKinsey Global Institute found interaction workers spend 28% of the workweek managing email and roughly 19% more hunting for internal information, and email is the medium that carries the entire quote cycle. McKinsey separately reports that companies expect digital procurement to cut time spent on transactional sourcing by 30 to 50%, and The Hackett Group's 2023 study put e-sourcing adoption at about 74% of organizations, which means roughly a quarter of large enterprises still run sourcing with no tool at all.

Running both: keep Excel, move the system of record

This is not a dual-run architecture; Excel is the starting point, not a companion system. What actually happens is migration without abandonment.

LightSource ingests the spreadsheets a team already has: drop an Excel file, PDF, or drawing and items are created or updated automatically, with flat-BOM templates for bulk loads, so clean data is not a precondition for starting. Suppliers keep quoting exactly the way they do today through Excel round-trip: download the bid sheet, fill it offline, upload it back. Every grid, bid sheet, and analysis exports back to Excel for whatever modeling comes next.

Excel does not disappear. Cost engineers will keep building cleansheets and one-off models in it, and they should. It goes back to being what it is great at, ad hoc analysis and scratchpad modeling, and it stops being the system of record for parts, quotes, and supplier history.

What public reviewers say

LightSource holds 4.9/5 on G2 across roughly 11 reviews; there is no G2 category for the status quo, so the closest proxy is Microsoft Excel itself, at 4.7/5 across more than 3,100 reviews. The caveat cuts both ways: the review bases differ by two orders of magnitude, and Excel's reviewers are rating a spreadsheet, not a sourcing system.

Excel's reviews are admiring. A G2 reviewer, a CTO (Mid-Market, 2026), writes: "What I like best about Microsoft Excel is its power and flexibility for complex data analysis. It offers advanced formulas, pivot tables, and strong performance when working with large datasets." The same review lists the cost of that flexibility: "Collaboration in Microsoft Excel can be limited and often leads to version conflicts when multiple people edit files ... Managing large shared files can also become messy and hard to track." Another G2 reviewer, an IT recruiter (Mid-Market, 2026), names the risk the error research documents, calling the grid "prone to human error," where "a small mistake in a formula, wrong cell reference, or accidental overwrite can impact the entire sheet."

On the LightSource side, a G2 reviewer, a VP of Supply Chain (Mid-Market, 2024), writes: "Having the cost and quoting information in a single location helps with rapid product launches and better visibility for our leadership group. It's got its quirks, but the constant updates are promising." And a G2 reviewer (Mid-Market, 2024) who deployed it across multiple portfolio companies makes the turnover point directly: "All the RFPs are centrally located, so no loss of data when a SM leaves the organization."

In verified enterprise reviews of sourcing tools generally, spreadsheets appear most often as the before state: the manual bid comparisons and scattered quote files reviewers describe being glad to leave behind.