Skip to main content

Which platforms are worth building on?

245 products were checked against six axes and thirty published checks, from primary sources only. This page shows that work two ways: the programs we would actually recommend building on, and the survey they were drawn from.

Programs worth partnering with

26 of the 160 qualifying programs make this list. A program is here when it scores 55 or better on the weighting below and it has a marketplace whose listings we could count. Publishing well while nobody has ever counted the catalog is not yet a distribution argument, so the second half is not a formality — it holds back 21 programs that clear the score alone.

Change the weighting and this list changes with it — that is the point of it being a weighting rather than a verdict. Everything that does not make the bar is still on the page, under The whole landscape.

This list is not the study’s sample. It mixes category leaders with competitors and requested additions, because a buyer choosing a platform does not care which of them was picked first. The correlation and coverage figures under The whole landscape are computed over category-leader rows only, and do not describe this list. Every row below says which it is.

26 programs. Click any row for the thirty checks and the pages we read. 24 of the 26 listed here carry a marketplace size we counted ourselves, not one read off a vendor landing page.

Partner programs ranked by weighted score, with per-axis scores and marketplace size
ProgramScoreListings
1. OneDriveConsumer Cloud Storagecompetitor86.588
2. BigCommerceE-commerce Platformscompetitor84.01,350
3. BigQueryData Warehouse / Lakehousecompetitor80.5156
4. GitHubCI/CD · Version Controlcategory leader78.01,301
5. WixE-commerce Platforms · Website Builderscategory leader77.555
6. KlaviyoEmail Marketing / ESP · Marketing Automation · 5 of 6 axes applyrequested addition77.2339
7. XeroAccounting / Financecompetitor71.51,233
8. HubSpotCRM · Marketing Automationcategory leader70.02,440
9. WooCommerceE-commerce Platformscompetitor69.01,455
10. AcronisBackup & Recovery · 5 of 6 axes applycompetitor67.5246
11. Google Workspace MarketplaceConsumer Cloud Storage · Email Clientscategory leader67.5109
12. Adobe CommerceE-commerce Platformscompetitor66.03,705
13. ShopifyE-commerce Platformscategory leader65.024,521
14. ZendeskHelp Desk / Customer Supportcategory leader65.01,758
15. GitLab CI/CDCI/CD · 5 of 6 axes applycompetitor64.441
16. QuickBooks OnlineAccounting / Finance · 5 of 6 axes applycategory leader62.7550
17. TwilioCPaaS / SMScategory leader61.523
18. SegmentCustomer Data Platform (CDP) · 5 of 6 axes applycategory leader61.3700(vendor)
19. ZapieriPaaS / Workflow Automation · 5 of 6 axes applycategory leader60.29,875
20. DatabricksData Warehouse / Lakehouse · 5 of 6 axes applycompetitor60.014
21. RingCentralVoIP Business Phone · 5 of 6 axes applycategory leader59.5568
22. DoorDashFood Delivery · 5 of 6 axes applycategory leader58.361
23. ZoomVideo Conferencing · Webinar Platformscategory leader58.03,596
24. SalesforceCPQ (Configure-Price-Quote) · CRMcategory leader56.55,455
25. Autodesk Construction CloudConstruction Managementcompetitor55.0400(vendor)
26. MailchimpEmail Marketing / ESP · 5 of 6 axes applycategory leader55.0341

Compare inside a category

Two programs in the same category face the same buyers and the same norms, so a difference between them means something. A difference between categories mostly measures the categories. 41 of the study’s categories hold more than one of the 160 qualifying programs, which is what a head-to-head needs.

How this was measured

Every score comes from a page we could open without a login, an NDA or a sales call. Where a check reads 0 it means we looked and it was not published — not that we could not find it. Where a check reads n/a the rubric does not ask that question of that program, and it is left out of the denominator rather than counted as a failure.

Transparency

Is the program legible without a sales call?

  • Guide public — The program guide/docs are readable with no login, no NDA, no lead-capture form. A form-gated PDF is 0.
  • Requirements numeric — Tier or listing requirements are stated as numbers (install counts, revenue thresholds, cert counts, SLAs) rather than adjectives
  • Reward numeric — The revenue share / commission / margin is stated as a number somewhere public
  • Rules of engagement — Deal registration, attribution, or conflict-resolution rules are published where partners can find them
  • Changes dated — Term changes carry an effective date announced in advance, or thresholds are published with validity windows

Demand

Does the platform bring you buyers, or just host you?

  • Lead share — The platform commits to routing leads or bringing you into deals. A directory listing alone is 0 — a directory is pull, not lead share
  • Shared pipeline — Account mapping, co-sell eligibility, or a shared deal record is offered
  • Co-marketing specified — Co-marketing exists and who qualifies / how it works is stated
  • Funding — MDF, credits, or co-marketing funds with a published path to access
  • Merchandising — Events, speaking slots, featured placement, or badges with stated criteria

Entry cost

What must you pay to be here? Higher is cheaper.

  • No entry fee — No paid membership required to join the program
  • No submission fee — No per-submission review/certification fee
  • Fast review — Review/certification typically ≤4 weeks, or there is no review
  • No annual fee — No recurring listing or membership fee
  • No paid tier gate — A basic public listing does not require a paid tier

Distribution

Can customers find you, and can you influence that?

  • Browsable directory — A public directory end customers actually browse (not a logo wall or a first-party connector list)
  • Earnable placement — Criteria for featuring / ranking / badges are published, so better distribution is earnable rather than granted
  • Analytics returned — The partner gets install, usage or listing-performance data back
  • Listing control — The partner controls category, search terms, media, or positioning
  • Audience disclosed — The platform publishes a real audience signal — listing counts, install counts, MAU, traffic

Economics

Can you actually transact — and on what terms?

  • Price visible on the listing — Someone browsing the listing can tell what the product costs — a price, a starting price, or a plan table on the listing surface itself
  • They collect and remit — The platform bills the end customer and pays you — you don't have to build billing
  • You set the price — The partner controls pricing (vs platform-set or fixed tiers)
  • Provisioning documented — What you must build to transact is documented — provisioning / entitlement / billed-on-behalf-of endpoints, webhooks, license checks
  • Payout terms published — Payment timing, minimum thresholds, currencies, processing fees

Terms

How onerous is the agreement? Higher is lighter.

  • No exclusivity — No requirement to be exclusive to this platform or to refrain from competing marketplaces
  • No broad IP grant — The platform does not take broad rights over your app, IP or derivative works
  • No clone reservation — The platform does not reserve a general right to build competing functionality, or explicitly commits not to remove apps for competing
  • Fair termination — Termination requires cause, or a notice period is published
  • You keep the customer — The partner retains the customer relationship and access to their own customer/usage data

The known weak points

  • Five programs turn on a judgement we could not read. Whether a listing can transact decides six of its thirty checks, and for four programs the evidence points both ways or the page would not load. We hold those at their scored value rather than guessing, and name them — guessing low would raise a score on evidence nobody has.
  • The price-visibility check is mostly unscored. Whether a buyer can see a price on the listing is answered for 5 of the 63 programs it applies to. The rest are marked not-assessed, not zero: a zero would claim the marketplace hides its prices, which is a statement about them rather than about us.
  • The weighting is an opinion — which is why you can change it above.
  • The shortlist is a judgement too. A score of 55 plus a countable marketplace is our bar, not an industry one, and it is applied to a survey built to span the market rather than to rank it. Read it as where we would start, not as a ranking of everyone.
  • Some sites blocked us. A citation marked blocked means the page exists and our fetch failed, so that score is likely understated. That is a different fact from an absence, and the program pages distinguish them.
  • Categories are editorial. A dozen of them mix products that do not really compete.

Compiled 2026-08-19. Scores are a reading of what a platform publishes, not a judgment of the platform. Method and raw data