Choice modelling & live market simulation

Clear thinkers take feedback from reality, not consensus.

Opinion surveys ask what customers would like. Ordeen measures what they actually choose when a real price is attached, fits a Bayesian preference model to those choices, and hands you a live simulator for every price, spec and competitor move still on the table.

Step 01 of 07Typical time: 2 mins

Start with a brief. Ordeen drafts the study.

Describe the commercial decision you need to make and attach any strategy deck or CSV. Ordeen extracts your product specs, target market, and price tiers, and recommends the right methodology out of six.

How methods are chosen →
UPLOADED BRIEFQ4 Pricing.pdfDECISION OBJECTIVE“Re-price Starter, Pro & Entand test bundling SAML SSO.”✦ 3 Plan Tiers Detected✦ 6 Core Features Found📎 Competitor_Prices.csv✓ EXTRACTED IN 12 SECONDS→★ RECOMMENDED FITChoice-Based Conjoint+ Conditional Tier PricingBest for pricing 3–7 attributes against competitorswhere each plan tier carries its own price ladder.ALSO AVAILABLE IN WORKSPACE:Adaptive Conjoint (ACBC)8+ attributesAnchored MaxDiff10–30 feature rankMenu-Based Choice (MBC)À-la-carte bundlesVan Westendorp & Gabor-GrangerFast single SKU
Working days to simulator2–4Design, panel, Bayesian estimation and simulator live in one platform. The only wait is respondents.
Countries with built-in panel18Integrated Prodege panel at ~£3–£4/complete with World Bank quotas, or survey your own list for £0.
Workspace subscription£10k/qtrOr £35,000/yr. 5 creator seats + 10 executive viewer seats on shared live simulators.
How a study runs

One workspace. Yours within days, not quarters.

01 · Day 1 (15 minutes)

You upload the brief.

Describe the decision and attach any deck or CSV. Ordeen recommends one of the 6 methods, drafts your competitor SKU table and attributes, and scores level balance & orthogonality before launch.

02 · Days 1–3

Real buyers trade off.

Launch to our integrated Prodege panel across 18 countries (~£3–£4/complete with World Bank quotas) or your own customer list. A 60–90s warm-up game and automated fraud checks clean responses live.

03 · Automatic (~10 mins)

Bayesian estimation & segments.

Multi-chain Hierarchical Bayes estimates individual utilities. Reliability-weighted k-means++ and Bayesian Latent Class discover and name your buyer segments.

04 · Ongoing

You simulate & run waves.

Test prices, bundles and TURF line-ups live. Share password-protected simulator links with 10 stakeholders, export respondent CSVs, and click Run next wave next quarter.

One strategy brief · Three ways to run it ↓
Route 01 · The Agency ProjectStatic PDF

Six hand-offs. One frozen deck.

PARTY / HAND-OFFWK 1WK 3WK 6WK 801 · YOUR TEAM02 · AGENCY DIR03 · PROGRAMMER04 · PANEL BROKER05 · STATISTICIANPDF60ppfinal_pricing_deck_v7.pdfCFO asks: “What if we charge £49 instead?”× DEAD ARTIFACT · CANNOT RE-QUERY

Your brief passes through five external teams over eight weeks. You pay £60,000 for a 60-page PDF that cannot answer a single new scenario once the boardroom meeting starts.

Time to answer6–8 weeks
Typical cost£50k–£100k / study
Route 02 · Survey-Suite Add-OnFragile Maths

A conjoint widget, bolted on.

GENERIC SURVEY TOOLBOLTED-ON ADD-ONWTP RATIO: £∞ (NO CI)MISSING ECONOMETRIC SAFEGUARDSERR 01Prohibitions correlate tier & price×ERR 02No R-hat / ESS convergence check×ERR 03Point-ratio WTP blows up (±∞)×

Bolting a basic conjoint block onto a general survey tool breaks on Good/Better/Best tier ladders, hides whether the Bayesian chains converged, and prints WTP ratios that blow up on noise.

Tiered pricingCollinear / broken
Convergence gateUnverified black box
Route 03 · Ordeen Workspace★ Live Model

One workspace. Brief to live simulator.

01 BRIEF02 DESIGN03 PANEL04 LIVE SIMR-HAT 1.008 · 95% POSTERIOR CIs★ OPTIMAL MARGIN · REV +28%★ Your Pro Tier27.4%+3.7ppCompetitor A26.8%-1.9ppCompetitor B20.6%-1.0ppWalk-away (None)14.4%-0.8ppLIVE PRICE SWEEP (CLICK):£35£39 ★£45

Your team goes from brief to built-in Prodege panel to an audited Hierarchical Bayesian simulator in days. Click any price step above to watch share, 95% credible intervals, and revenue update live.

Time to live model2–4 days (fieldwork)
Workspace licence£10k/qtr or £35k/yr
Where the tool takes you

Four questions every pricing decision turns on.

How each option handles them:Answers itGives you part of itDoesn't answer it· Detailed comparisons: vs Sawtooth · vs Qualtrics · vs Conjointly · vs quantilope · vs QuestionPro · vs SurveyMonkey

01. "If we raise our Pro tier from £39 to £49 and add SSO, do we gain revenue or lose share?"

Read: Conditional pricing for tiers →
Generic Survey Suite (Qualtrics / Alchemer)Doesn't answer it
Testing Starter (£19) and Enterprise (£399) on one price list creates £399 Starter cards. Blocking them with prohibitions correlates tier and price, breaking the model.
60-Page Agency Deck / Desktop SoftwareGives you part of it
Tests the 2 or 3 static scenarios you thought of six weeks ago. Testing a new price point or unit-cost margin requires going back to the analyst.
Ordeen Live SimulatorAnswers it
Designs on orthogonal price steps, fits in tier-centred log-price, and lets you drag price live across 25 rungs with paired-draw 95% credible intervals.
Illustrative From an example SaaS studyPro £39 → £49 (+SSO): Share moves 22.4% → 26.1% (Paired Δ +3.7pp, 95% CI +2.4 to +5.0pp). Net revenue +34%.

02. "How much is each feature upgrade actually worth in money, and can we trust the ratio?"

Read: Why naive WTP formulas lie →
Direct Survey Question / Van Westendorp AloneDoesn't answer it
Measures stated price resistance on one fixed product in a vacuum, ignoring competitor switching and individual feature value.
Standard Conjoint Point-Ratio WTPGives you part of it
Divides average feature utility by average price slope. When price sensitivity is small or heavy-tailed, Jensen's inequality makes the ratio explode.
Ordeen Dual-Lens WTP EngineAnswers it
Computes convexity-preserving Krinsky-Robb intervals anchored to your focal product, plus the exact share-holding price move against competitors. Refuses degenerate slopes outright.
Illustrative Example WTP tableSAML SSO: Focal-buyer valuation +£14.80 (±£2.10) · Competitive share-holding price move: +£11.50/mo.

03. "Which 3 SKUs or roadmap features reach the largest unduplicated share of our market?"

Read: Competitive Portfolio TURF →
1-to-5 Rating Scales / Kano SurveysDoesn't answer it
Every feature is rated 4.2 out of 5. Picking the top 3 items often selects 3 features that appeal to the exact same 40% of buyers.
Legacy Standalone TURFGives you part of it
Counts a buyer as reached if your SKU beats buying nothing, ignoring rival brands on the shelf and falsely predicting 99% reach on 2 SKUs.
Ordeen Unified Competitive TURFAnswers it
Measures probabilistic reach against both named market competitors and the walk-away ceiling, stopping at the exact ≥2pp statistical elbow.
Illustrative Example CPG range review3 SKUs capture 78.4% (±2.1%) of the 86.0% category ceiling. 4th SKU adds +1.2pp (below 2pp elbow, cut from range).

04. "Did the Bayesian model actually converge on clean human data?"

Read: R-hat, ESS & Divergences →
Black-Box Cloud ConjointDoesn't answer it
Hides MCMC chain diagnostics. If the sampler failed to mix or bots sped through the survey, you still get a confident-looking chart.
Manual Statistician WorkflowGives you part of it
Requires exporting raw CSVs after fieldwork to clean speeders by hand and run multi-seed convergence checks externally.
Ordeen Fail-Closed Estimation GateAnswers it
Runs a 60–90s warm-up game and AI open-end fraud screen during fieldwork, then checks R-hat, ESS, divergences and holdout hit rate. Non-converged fits are withheld.
Illustrative Diagnostic strip on every studyR-hat 1.008 (Pass) · Min ESS 640 · Divergences 0 · Bootstrap-CV Holdout Hit Rate 79.4% (±1.6%).
Interactive specimen · Try the simulator

Test a live pricing & packaging scenario right now.

Illustrative Click a price rung or toggle a feature below. The numbers shown are from a fabricated example SaaS scenario and are for interactive demo only. In a real Ordeen study, these numbers recompute from your fielded respondents' individual Hierarchical Bayesian posterior draws.

How the simulator works →
Illustrative Example Study: B2B SaaS Packaging (CBC with Conditional Pricing)Segment Filter: All Respondents (100%) · Paired CRN Draws Active
Illustrative Revenue Index (Per 1,000 Prospects)£8,736 / moBaseline
Posterior-Expectation Share of Preference (with 95% Credible Intervals)
★ Your Pro Tier (£39 · Standard)22.4% (20.8–24.0%)
Your Starter Tier (£19 · Fixed)18.0%
Incumbent Competitor A (£55 · SSO included)31.2%
Challenger Competitor B (£25 · Basic)14.0%
Walk-away (Choose None)14.4%
What the posterior draws tell you

At £39 without SSO, you lose enterprise switchers to Competitor A.

Try clicking £49 / mo and + SAML SSO Included above. Because SSO commands a +£14.80 focal-buyer WTP, adding SSO while raising price by £10 actually increases share by +3.7pp and lifts revenue by +34% in this illustrative scenario.

Point Elasticity at Priceε = -1.12 (Unit-elastic)
Pr(Pro Wins vs. Baseline)Baseline
First-class mathematical rigour

Three commitments. All of them checkable.

We don't ask you to take our numbers on trust. Every diagnostic is printed on the result, every safeguard is automatic, and our entire statistical methodology is published openly.

01 · We refuse before we guess

Enforced convergence & slope refusal

An MCMC sampler always outputs numbers, even when chains never mixed. Ordeen checks split-chain R-hat, Effective Sample Size (ESS) and divergences on every fit, withholding any model that fails to converge, and refusing any WTP figure whose price slope changes sign across draws.

02 · No Representative Consumer Fallacy

Posterior-expectation shares & Shapley

Plugging average utilities into a simulator washes out polarized niches, and traditional range importance changes whenever you widen the price ladder. Ordeen integrates shares across every respondent and posterior draw, and computes level-invariant Shapley importance.

03 · Multiple-testing & noise protection

Signal-weighted segments & FDR control

Standard k-means lets noisy attributes drag cluster centroids around, and testing 50 attribute interactions guarantees false positives by chance. Ordeen weights k-means by hierarchical reliability, runs Bayesian Latent Class alongside it, and gates interactions with Benjamini-Hochberg FDR.

The complete reference library · 30+ guides

Pricing and product research, with the maths in front of you.

Plain-English guides for strategy, product and pricing teams at every step from picking a method to defending the simulator in the boardroom.

View all guides on /learn →
Choose your workspace

Less than the cost of one consultancy study. Unlimited studies all year.

Both plans give your team the complete Choice Modelling platform, with all six methods, Hierarchical Bayes and Latent Class estimation, the live simulator, and 10 executive viewer seats on shared links. Respondents are bought at cost (~£3–£4/complete) in non-expiring Prodege packs, or fielded to your own customer list for free.

Quarterly Workspace · No annual lock-in

Quarterly Plan

£10,000Per quarter · 5 creator seats + 10 live simulator viewers
  • All 6 methods: CBC, ACBC, Anchored MaxDiff, MBC, Van Westendorp & Gabor-Granger
  • Unlimited studies and longitudinal tracking waves during the quarter
  • Hierarchical Bayes & Bayesian Latent Class with convergence diagnostics
  • Live Market Simulator with 95% credible intervals, WTP, 2D heatmaps & TURF
  • Respondent-level CSV, XLSX, JSON & Parquet exports + segment typing tool
Book a 20-minute demo →
Annual Workspace · Save £5,000 a year

Standard Annual Plan

£35,000Per year · 5 creator seats + 10 live simulator viewers
  • Everything in the Quarterly Workspace for the full 12 months
  • Saves £5,000/year compared to quarterly billing
  • Continuous multi-wave tracking as competitors and market prices move
  • EU-hosted in Google Cloud Belgium; DPA included; independent security reviews (April & October 2026)
  • Single annual invoice for procurement; extra 5-user packs at tier rate
Book a 20-minute demo →
Retail Intelligence · £2,000 / mo

Price Monitoring →

Track real retail checkout prices (price + delivery), promos and variant stock across 200+ retailers in 18 countries with screenshot evidence.

Read guide →
Supply & Finance · £2,000 / mo

Demand Forecasting →

AI proposes external signals; econometrics validates them out-of-sample. Includes Croston/TSB intermittent demand and MinT reconciliation.

Read guide →
Legal & Procurement · £1,000 / mo

Contract Analysis →

Analyse agreements against governing-law statute files and your company role; score portfolio risk, consolidate MSAs and draft renewals.

Read guide →
Verified data first, AI last

Bring your brief. We'll draft the market live in 20 minutes.

Walk through a live simulator with the team that built the estimation engine, and see how your pricing or product question maps to a study.

Book a 20-minute demo →
Or email im@ordeen.com