All Case Studies
Four claims we stand behind
Every claim below has a bold form and a conservative form. We use the bold form when the audience is close enough to verify; we use the conservative form on cold outreach so we under-promise by design. Either form is defendable against this page on the date stamped at the top.
Claim 01
As of 2026-09-13
88% of paying-client revenue is repeat business
88
percent
Public form
Bold: 88 percent. Conservative public form: 60 percent or more.
Definition + data
Counted as: paying customers with two or more paid invoices, share of total paid revenue from that cohort. SQL pull against the HRMS production invoices table on 2026-09-13. 199 paying customers total, 132 repeat, Rs 1.42 Cr total paid, Rs 1.26 Cr repeat (88.6 percent). The 2026-06-01 pull on the same query gave 84.4 percent.
Source
Query: Team AI repo: data/queries/repeat-revenue-2026-q2.sql
Audit doc: Team AI repo: posts/2026-06-01-dcrayons-defendable-numbers-sources.md (section 4)
Claim 02
As of 2026-09-13
10 years operating
10
years
Public form
Bold: 10 years. Conservative form: a decade.
Definition + data
Dcrayon Consultancy Private Limited registered 2016. Counting completed calendar years through 2026-09-13 gives 10. The Dcrayon Inc. US entity was registered later; the headline counts the Indian parent.
Source
Query: MCA registration filing (year of incorporation 2016)
Audit doc: Team AI repo: posts/2026-06-01-dcrayons-defendable-numbers-sources.md (section 2)
Claim 03
As of 2026-09-13
44-person team across functional departments
44
people
Public form
Bold: 44 active employees.
Definition + data
Active employees in the HRMS users table on 2026-09-13 (is_active = 1 with an employee code): Marketing and Growth 13, Design and Creative 11, Sales and Business Development 6, Website Development 5, Operations 2, People and HR 2, Client Success 2, unassigned 3. DEFINITION CHANGED: the 2026-06-01 figure of 88 counted every user row not deleted, which by September also included 66 former staff whose records were never closed. That definition would now print 115, so it was retired rather than inflated.
Source
Query: HRMS users table on AWS 144 (live)
Audit doc: Team AI repo: posts/2026-06-01-dcrayons-defendable-numbers-sources.md (section 1)
Claim 04
As of 2026-09-13
617 client records; 196 have paid
617
client records ever
Public form
Conservative roll-up: 200 or more paying clients across India, UK, US, UAE. Specific cuts: 617 client records opened; 196 have paid at least one invoice; paying clients by year 37 (2024), 55 (2025), 119 (2026 to date). The 'active clients' figure elsewhere on this site is 485, the CEO-confirmed count of the CRM client book (crm.clients, status Active) on 2026-06-21. On this site it is a fixed figure, not a live read.
Definition + data
From the HRMS customers and invoices tables on 2026-09-13: 617 customer records not deleted; 196 customers with at least one invoice paid or part-paid with a paid amount above zero; paying customers by year of payment 37, 55 and 119. The 2024 and 2025 figures match the 2026-06-01 pull exactly. Earlier versions of this page said the active-clients number was read live on every render; since the move to the current CMS it is not, so that claim was withdrawn. An earlier sub-figure of 79 'on an active paid retainer' was withdrawn too: it counted active accounts, not retainers.
Source
Query: crm.clients (live) + HRMS invoices (historical)
Audit doc: Team AI repo: posts/2026-06-01-dcrayons-defendable-numbers-sources.md (section 3)
How we refresh these
The four queries above re-run on a fixed quarterly cadence. Next scheduled refresh: 2026-10-15. If a number drops below the published bold figure at the next refresh, the bold figure on this page comes down within 48 hours of the pull and the conservative public form takes over until the next quarterly window.
If you find a public Dcrayon claim that does not match this page, tell us. Email info@dcrayon.com with the URL and the figure; we either fix the public copy or open the source query for you.
The discipline behind these numbers
Source-linked claims only hold if the words around them meet the same standard. Three pages document how we keep the copy honest.
The word ban
42 words we never print
The full ban wall: every banned word struck through, with the plainer replacement we use instead.
The refusals
What we refuse to print
Claims we turned down because the source data could not back them, listed by category.
The prompt
The writing prompt we use
The exact prompt our writers work under. Copy it and run it on your own content.
Want a number we have not published yet?
On a discovery call we will share the cuts you actually care about: managed Meta spend by vertical, repeat-cohort retention by tenure, share-of-answer by AI engine. We do not publish them because they are buyer-specific.