MBA
Compare

MBA vs Constructor.com

Enterprise AI commerce platform. AI Shopping Agents, Reasoning Engine, named Merchant Intelligence Agent. Typical contracts $150-$300K/year.

Platforms compared:Multi-platform
1Constructor is enterprise. $150K to $300K typical annual contract, multi-stage procurement, multi-month integration. Their logos are Sephora, Under Armour, Gap, REI, Petco, QVC, Williams-Sonoma.
2MBA is mid-market: $49/month flat, no usage caps, no rev share, self-serve. Same inspectable-rules outcome that Constructor sells as 'Merchant Intelligence Agent' for two to three orders of magnitude less.
3Constructor's 'Reasoning Engine' positions transparent rules as outdated. In practice the audit answer to 'why this recommendation?' is always rule-shaped: this pair shows up at this support and lift in this many baskets. We treat rules as the explainability surface, not the legacy.
4MBA's six MCP tools turn the same recommendation layer into agent-callable infrastructure for Claude Desktop, Claude Code, Cursor, Cline, and the OpenAI Agent SDK. Constructor's AI Shopping Agents are vendor-hosted, not agent-callable.
5If you're a mid-market VP eCom who got priced out of Constructor, the price-arbitraged answer is straightforward, with no capability lost on the rules-transparency axis.

When MBA wins

You're a mid-market merchant who was just quoted $250K by Constructor and can't get budget, but you still want inspectable rules + a substitution engine + agent-callable MCP tools.

When Constructor.com wins

You're a $100M+ GMV enterprise with budget for a Reasoning-Engine-class platform, a procurement process that prefers named brands, and a multi-quarter integration timeline.

Feature comparison

Color-striped rows favor MBA, Constructor.com, or are even. We mark each so you can scan for the trade-offs that matter to you.

FeatureMBAConstructor.com
Pricing band
$49-$199/month flat, published on the homepage, no usage caps, no rev share.
$150K-$300K/year typical contract per public case studies and analyst reports. Quote-only, multi-stage procurement.
Inspectable recommendation logic
Every rule shows support, confidence, lift. Merchandiser can override or suppress per opportunity. The 'why this recommendation?' answer is in the admin grid.
Branded 'Merchant Intelligence Agent' surfaces ranking explanations. The capability is real but reserved for high-tier contracts.
Reasoning vs rules framing
We mine real co-purchase rules from your order history. Rules ARE the audit trail. We treat 'reasoning beats rules' as marketing, not engineering, because in commerce the question 'why did you recommend this?' has a concrete data answer.
Markets 'Reasoning Engine' as the AI primitive. Positions rules-based systems as outdated. The actual mechanics are not public.
AI Shopping Agents
MBA exposes recommendations + substitutes + quote bundles as MCP tools any agent host can call (Claude Desktop, Claude Code, Cursor, OpenAI Agent SDK). Tools are the agents.
Named 'AI Shopping Agents' product, fronted by their own conversational UI. Vendor-hosted agent layer rather than agent-callable tools.
Time to first value
10-15 minutes after install (run a mining job, see rules in the admin).
Multi-month implementation typical, including catalog ingestion, taxonomy mapping, and trained-model warmup.
Multi-platform support
Shopify, WooCommerce, Magento / Adobe Commerce, OroCommerce. Same rules layer across all.
Headless / API-first. Integrates with most platforms but optimized for enterprise-custom storefronts.
B2B / wholesale
Native MCP tool propose_quote_bundle, account-aware mining, RequisitionList integration on Adobe Commerce B2B.
Enterprise B2B available as part of larger contracts. Not packaged as a separate SKU.
Customer reference logos
Early-stage. We're building case studies from beta merchants and will publish them as they land.
Sephora, Under Armour, Gap, REI, Petco, QVC, Williams-Sonoma. Strong enterprise reference list.
Procurement fit
Self-serve. $49 plan + Stripe = you're paying in a minute.
Quote, SOW, security review, MSA, multi-month procurement cycle. Standard enterprise SaaS process.

Pricing snapshot

MarketBasketAnalysis

Flat $49/mo for WooCommerce, $199/mo for Magento, $49/mo for Shopify (in App Store review). No usage caps, no revenue share, no add-ons.

See plans

Constructor.com

Quote-only enterprise contracts, typically $150-$300K/year. Customer logos include Sephora, Under Armour, Gap, REI, Petco, QVC, Williams-Sonoma.

Visit Constructor.com

Pricing at GMV tiers

Rough cost comparison at representative store sizes. Real numbers from real procurement land within a wide band; this is the math we'd walk you through if you emailed and asked.

TierMBAConstructor.com
$5M GMV / mid-market$49/mo flat ($588/year)
Constructor entry contract typically starts in the $100K+/year band even for smaller mid-market
Roughly 170x cost difference at this band. Most mid-market merchants cannot get the Constructor quote past finance.
$25M GMV / upper-mid-market$49/mo flat
Typical Constructor contract $150-$200K/year
Constructor's strongest fit band starts here. They invest serious implementation services per account.
$100M GMV / enterprise$49/mo flat (or hosted enterprise on request)
Typical Constructor contract $200-$300K/year
At this scale the comparison is honestly different. You're probably already in an enterprise RFP, and Constructor is on the shortlist. MBA's role here is closer to 'agent-callable data layer' than 'primary recommendation platform.'
$500M+ GMV / Sephora-tierHosted enterprise pricing on request
Enterprise contract; Constructor's reference logos cluster in this band
Different procurement entirely. Constructor's Reasoning Engine + implementation services are sized for this band; MBA is not pitched here.

When to pick each, expanded

The above-the-fold version is one sentence each. Here are the full bullet lists.

Pick MBA when

  • You're $5M-$50M GMV and got a $150K+ quote from Constructor that finance won't approve.
  • You want inspectable rules (support / confidence / lift) but you're not paying $200K/year to get them.
  • Your AI strategy is agent-callable tools (MCP) more than vendor-hosted shopping agents.
  • You want native multi-platform support (Shopify, WC, Magento, OroCommerce) instead of headless-API-first integration work.
  • Time to first value matters: 15 minutes to a populated admin grid, not a 12-week implementation.

Pick Constructor.com when

  • You're $100M+ GMV with budget for an enterprise-class platform.
  • Your procurement prefers named enterprise brands with deep reference customer lists.
  • You have engineering capacity for a multi-quarter integration including catalog ingestion + taxonomy mapping.
  • You want a vendor-hosted AI Shopping Agents conversational layer (rather than wiring your own via MCP).
  • Your competitive set is Sephora / Gap / Williams-Sonoma and you need procurement signaling that matches.

The honest verdict

1

Constructor is real enterprise infrastructure. If you're at Sephora scale, the comparison is genuinely a different procurement: you're buying a Reasoning-Engine-class platform plus implementation services, and MBA is not in that lane.

2

If you're a $5M-$50M GMV merchant who was just quoted $200K+ by Constructor and can't get budget, MBA exists exactly for you. Inspectable rules, substitution engine, MCP tools, and a $49 monthly bill that does not scale with GMV.

3

Constructor's 'Merchant Intelligence Agent' and MBA's inspectable rules are the same idea: show the merchandiser why a ranking decision was made. The difference is procurement fit, not capability.

4

The 'reasoning beats rules' framing Constructor uses is marketing. In commerce, the audit answer to 'why did you recommend this?' IS a rule (this pair shows up in 1,200 baskets at 4.2x lift), so transparent rules + reasoning are not opposed, they're the same thing told two ways.

5

Pick Constructor if you need the enterprise reference logos and have budget. Pick MBA if you want the same inspectable-rules outcome on a flat $49 bill.

Frequently asked, MBA vs Constructor.com

Is Constructor really $150K-$300K per year?

That's the typical contract band per public analyst reporting and procurement leaks. Smaller mid-market accounts sometimes come in lower; enterprise accounts go higher. The published pricing is quote-only, so the public numbers are estimates.

What is Constructor's Merchant Intelligence Agent and does MBA have an equivalent?

Constructor's MIA shows merchandisers why a given ranking decision was made (the underlying signals + score components). MBA's inspectable-rules admin grid is the same concept: every recommendation shows support, confidence, lift, and the merchandiser can override or suppress it. We shipped the capability; we just don't brand it as an Agent.

What is the 'Reasoning Engine' framing?

Constructor markets a Reasoning Engine as the AI primitive that distinguishes them from rules-based recommenders. We think that framing is more marketing than engineering: the audit answer to 'why this recommendation?' is always rule-shaped, and treating rules as the explainability surface rather than the legacy is a feature, not a regression.

How does Constructor handle B2B?

B2B is available as part of larger enterprise contracts but not packaged as a separate product surface. MBA ships propose_quote_bundle as a first-class MCP tool with account-aware mining and a Magento + Adobe Commerce B2B RequisitionList integration on day one.

I'm at $20M GMV. Will my CFO approve Constructor?

Probably not without a serious revenue-lift business case and a multi-year amortization model. Most $20M GMV merchants either pay 4 figures elsewhere (Glood, Rebuy, Amplify) or pay $49 to MBA. The Constructor band gets hard to justify until your competitive set is enterprise.

Can MBA scale to $100M+ GMV?

The architecture scales: mining runs in async jobs, the storefront fetch is sub-millisecond against a materialized rules table, and we run on real production Magento and Shopify infra. Above $50M GMV most merchants want a hosted enterprise option (managed mining, SLA support); we offer that on request rather than as a self-serve tier.

What about agent integrations?

MBA's MCP server exposes get_recommendations, find_substitutes, get_bundle_for_cart, propose_quote_bundle, score_cross_sell, and predict_reorder as agent-callable tools. Constructor's AI Shopping Agents are vendor-hosted conversational UIs. Different surface: tools an agent calls vs an agent the vendor hosts.

Should I run both?

Rarely. Constructor and MBA both occupy the recommendation-platform slot; running both adds operational complexity for marginal gain. The decision is usually procurement-driven: enterprise procurement picks Constructor, mid-market procurement picks MBA.

Want the full buyer's deep-dive?

This page is the per-vendor short version. For the long version, read “MBA vs Bloomreach” (8 pages on the enterprise-vs-mid-market trade-off) or “MBA vs the widget cluster” (4 pages of pricing math vs Glood, Rebuy, and PickyStory).

Try MBA on your own store

14-day money-back guarantee on all paid platforms. WP.org free tier on WooCommerce. No credit card required for the free path.