Recense vs SPSS

What stays the same

• The .sav file is the unit of work. Recense reads SPSS .sav natively — variable types, value labels, user-missing codes, weights, and multi-response definitions all round-trip.
• Survey-aware semantics. Bases honour user-missing codes. Weighting is first-class. Significance testing is built in.
• Crosstab as a primitive. Cross-tabulation, percentages, sig tests, and weighted bases work the way you expect. The numbers reconcile.

What changes

• Drag-and-drop, not dialog stacks. Build tables by dragging variables to rows and columns. No nested modal windows, no "Statistics → Custom Tables → Compute…" trees.
• No syntax window for power use. Derived variables, recodes, and formula cells live in the canvas. You can still write expressions when you want to — but nothing requires it.
• Tabulation in milliseconds. Rust and WebAssembly under the hood. Sub-second crosstabs even with 20 million+ data points — no progress bar, no waiting.
• AI agents in the same canvas. Brief an agent to build a brand-tracking grid by region with sig tests at 95%. It works in your workspace; the result is a real, editable table, not a screenshot.
• No round-trip to Excel. Formula cells inside tables compute net scores, indices, and point differences without exporting first.
• Browser-based, no install. Tabulation runs client-side via WebAssembly. Local files stay local. No Windows-only desktop application.

Cost

SPSS Statistics Standard sits well above £1,000 per user per year once you add the modules most researchers actually use. Recense Starter is £30 / month with full tabulation and BYOK AI access. For a five-person team that's roughly a £4,000 / year line-item difference at the same level of capability. See the pricing page for the current tiers.

Migration

Drop a .sav file into Recense and the dataset is ready in a few seconds. Variables, labels, missing codes, weights, and multi- response sets are all preserved. There is no separate dataset preparation step before you can build tables.

For agencies running shared scripts in SPSS syntax, those don't port directly today. The agent can rebuild most of what a syntax file does (recodes, derived nets, table specs) on demand — often with less work than translating the script — but for very heavy syntax estates this is a real migration cost. Worth a conversation if it's your situation.

Where SPSS still wins

Some advanced statistical modules in SPSS aren't yet first-class in Recense — full mixed-effects modelling, advanced regression diagnostics, and some specialised psychometric methods. These are on the roadmap, but if your work depends on one of them today, be honest with yourself about the timing.