●●

Cloudmaven — six weeks of work, where the strategy really stands, and what needs deciding

Monday 17 August 2026 · one review surface, commentable everywhere (select text or use the note tool — comments are stored centrally and any single chat can be pointed at one)

How this page was built, and what to trust. Every statement of fact here is sourced from the call transcripts — Lars 10 Aug, Christoph 13 Aug, the marketing team meetings of 10 Jul / 7 Aug / 14 Aug, the kickoff 8 Jul — or from a clickable artifact you can open and judge yourself. The planning and strategy documents in circulation are AI-written and unreviewed; they are treated as input, never as fact, and are marked UNREVIEWED wherever referenced. Transcripts themselves can be imperfect; where something matters, the wording is quoted so you can check the source.

1The job — and the bottleneck, in the client's own words

Cloudmaven pays for marketing that brings clients. The bottleneck is on record, from Lars (10 Aug): install-base business runs well and makes decent money — "was uns fehlt, ist die New-Business-Seite … über die Webseite passiert das bei uns leider gar nicht" — new companies simply do not find them through the website. His priorities, verbatim in order: 1) website, appearance, SEO, MQLs ("damit steht und fällt alles am Marketing"), 2) feed the install base, 3) content for outbound. His operating instruction: "mach das einfach. Wir haben großes Vertrauen in dich. Wichtiger ist, dass wir zeitnah erste Ergebnisse sehen und Gas geben." September is high season; he wants things standing by then.

Everything below is measured against those three pillars.

2The six weeks — what actually happened, week by week

When	What happened	Evidence
8 Jul	Kickoff with Lars: operating model, team and access, sales-material needs; transcript infrastructure set up so every call becomes a source.	Kickoff transcript
10 Jul	Marketing team met (Heidi content/coordination, Abdel web, Taimor SEO, Amr design/social, Inneke proofreading — team in Egypt, works in English).	Team meeting transcript
12–13 Jul	The Landing Page Lab built — the first complete offer set (AI-Specify, Cloudmaven × Everest): chooser + three full variants + drafts — then independently audited (two blockers found, both since fixed). ClickUp access/export calls with Lars + Jarrid, Eric.	the lab, live
26–28 Jul	First full site rebuild pass on the new stack; review PDF of the lab sent to Lars (28 Jul).	the rebuild, live
3 Aug	Lars call: short-term test agreed — a WordPress clone at dev.cloudmaven.com as the sensible quick win while the real migration was prepared.	3 Aug transcript
7 Aug	Marketing standup (Carsten leading): staging already ~50% faster (Abdel + Werner); German NetSuite-implementation page hit Google top-3 for six days; two team landing pages (NetSuite Partner Deutschland, Salesforce) content-complete, waiting on Lars's review; four success stories in progress (Stove, ZIS, TVEO, ShiftMove); Amr's Claude workflow producing creative briefs in ~10 min; content-strategy session with Heidi (white paper, fireside chats, Trovarit/university route, filming plan).	7 Aug standup notes
10 Aug	Lars strategy call — his strategy in his own words (§3), his design read (clean, bright, BearingPoint as reference, robots out, real people in), the migration green light ("mach das einfach"), 2–3 weeks to live discussed, his side: Cloudflare/GitLab rights + higher automation approval, Jarrid escalation if needed, November team-event photos, Heidi reprioritization.	10 Aug transcript
13 Aug	Christoph call — his positioning (§3), the first testable design rule anyone gave (1s / 3s / 30s), the world-map requirement (Middle East, South Africa, Australia, USA — not Germany-first), install-base goal (break-even from install base alone), the BDR contract signed that morning (starts 1 Sept), Knowledge Sales plan, faces UK = Fares / FR = Eric, 27 Aug in-person demo day fixed, 11:00–16:30, filmed interview planned, Jan identified as sole ultimate veto.	13 Aug transcript
14 Aug	Marketing team sync: AI-enablement plan for the whole team (everyone keeps their job, everyone gets better tools), note-takers standard, Claude Code rollout started, migration explained to the team, locations-page proposal (map first, then Lars+Christoph verify).	14 Aug sync notes
15–17 Aug	The build sprint: the exact copy proven word-identical (286/286 pages, twice, independently) · the theme engine (four full-site designs over one frozen copy source — design change is one setting) · v1.1 image modernisation (two variants) · v1.2 animated hero · working contact forms against the real HubSpot account · the address-shape launch blocker found, fixed, negative-tested · security/SEO/consent production checklist built · claims register audited · admin interfaces built · review + commenting infrastructure (this page) · access protection on every surface after two real exposures were found and closed.	§4 below — all clickable

3The strategy — honestly, from the two owners' own mouths

There is not one agreed strategy. There are two on record, they overlap on the goal and differ on the story. Anything that presents "the Cloudmaven strategy" as settled is ahead of the facts. This is the actual state:

Lars, 10 Aug — process-first, ERP as the door. "We position as the business-process transformers … system-diagnostic, not system-specific. Process first, tool/ERP as the vehicle second, AI agents and our own products third." ERP is the entry because "over ERP we own the whole IT landscape." Target: 70/30 ARR-to-services split, long-term ("ein Wunschgedanke"). Missing piece by his own account: industry specialisation — Software, Retail/E-Commerce, light manufacturing with software. Germany leads, other markets adapt. Small circle decides (Lars + Christoph + Carsten). On putting all four product areas on the site: "aus meiner Sicht keine gute Idee" — ERP focus, the rest as side doors.Source: 10 Aug transcript, his words

Christoph, 13 Aug — the CFO's whole surface, from one hand. ERP and EPM co-equal on top (the site serves Europe and the Middle East, which is EPM-led and does more revenue), CRM and HCM visible below: "Wenn wir als Trusted Advisor des CFOs auftreten wollen, müssen wir alle Themen abdecken können." The durable story: knowledge depth — "die einzige Story, die ich nach draußen gelten lasse, ist: Wir sind ein Multisystemhaus." Install base first: the goal is break-even from install-base business alone; a dedicated BDR starts 1 Sept; Knowledge Sales follows. Sales org toward 80–90% margin on ARR; services toward 35%.Source: 13 Aug transcript, his words

Where they agree — solid ground to build on

ARR is the goal. Both, independently, in almost the same words. Company value comes from recurring revenue, not services volume.
The install base is the nearest money. Lars: it "runs well" and needs feeding (his pillar 2). Christoph: break-even from it alone, with a dedicated caller from 1 Sept.
The website is the front and it currently produces nothing. Both said it; Christoph added that the current contact CTA produced zero enquiries in eighteen months.
Speed over deliberation. Lars: "Gas geben," results by September. Christoph: decides orally, from demos, fast.

Where they differ — the open strategic question

STRATEGYS1 — ERP-first (Lars) vs. four-areas-co-equal under CFO advisory (Christoph). Directly opposite, both on record, three days apart. Christoph is above Lars and stated his version twice including as a parting instruction — but Lars runs sales and will present whatever the site says. Carsten's own recorded position (10 Aug, to Lars): process-positioning is weak ground — "70 % der Konkurrenten sagen auch, wir sind die Prozessexperten" — alternatives should be examined together. This is the single most important unsettled question and it belongs in the room: 19 Aug with Lars, decided 27 Aug with Christoph. The site structure can carry either (the IA is data, not code) — but the story must be one story.

STRATEGYS2 — What is genuinely decided vs. what is AI-paper. Decided on the record: the design acceptance rule (1s/3s/30s, Christoph), faces (UK Fares / FR Eric), T&E off the site, install-base first, the map requirement, 27 Aug demo day, filmed interview, BDR from 1 Sept. NOT decided anywhere on record: the positioning sentence, industry focus, per-country product lists in writing (Christoph corrected himself mid-list — he asked for it in writing), EBIT vs EBITDA for the €4m, Health Check price, the qualified-lead definitions. The 40-page strategy document and "The Plan" UNREVIEWED contain useful analysis but nothing in them is a decision until one of the three of you says so.

4Everything that exists — clickable, labeled by what it demonstrably is

Artifact	What it is / what it proves	Pillar
cm-v1-0 — the exact copy	Today's entire site on the new engine, 286/286 pages word-identical (verified twice, independently). Proves the migration loses nothing. Loads instantly.	1
The theme engine	Four complete site designs over one frozen copy source — baseline, modernised, Ledger Light, Night Blueprint. Changing design = one setting; a gate makes changing a single word impossible. This is the machine that makes every future design change cheap.	1
cm-v1-1 · cm-v1-2	The copy with images modernised (two variants to pick between) · the copy with the homepage graphic replaced by the animated as-it-runs-today/after diagram.	1
Design directions, side by side	All design work in one place: the six lanes plus the full gallery — the photo-integrated homepage mockups, the Claude Design export (e-flow etc.), iterations r1–r3, early studies. Lars's stated taste (10 Aug): clean, bright, real people, robots out — judge against that and the 1s/3s/30s rule.	1
The Landing Page Lab	The landing-page machine with its first offer (AI-Specify): chooser + variants A/B/C + drafts. Audited, blockers fixed. The team's WordPress landing pages (NetSuite Partner DE, Salesforce — content-complete since 7 Aug, waiting on Lars's review) are the parallel track until cutover.	2·3
The claims register	Every public number with a source and an owner; feeds automated build gates. The tool that stops site, decks and proposals contradicting each other.	all
Copy directions	Three homepage copy directions (keep / pain-first / advisory-depth) as drafts, all watermarked NEEDS-VOICE. Raw material for the S1 decision — not proposals to ship.	1
Production readiness	Built and verified on branches: absolute canonicals/hreflang/sitemap, all 8 security headers with strict CSP, zero third-party requests before consent, redirect map (404s 209→15), the address-shape fix merged and negative-tested. The team's SEO substance (Taimor's audits, the top-3 ranking win) carries over 1:1.	1
Admin interfaces	Four surfaces (overview / SEO / content / install-base) reading real repo data or showing NOT CONNECTED — never an invented number. Needs a gated host to be clickable (C3).	2

5What we tell Lars today — against his own three pillars

Pillar 1 — website/SEO/MQLs: "The migration you green-lit on the 10th is real and provable: the whole site runs on the new engine, word-identical, radically faster — and design is now interchangeable, so your 'clean and bright, real people' direction and Christoph's requirements can be shown side by side instead of argued about. The go-live path is staged and reversible; the only external blocker is the Cloudflare/GitLab access you offered to unblock — that's your item from the 10th, plus the Jarrid escalation if he sits on it." Show: v1.0 next to the live site on a phone, then the themes.
Pillar 2 — install base: "Christoph's BDR starts calling the install base 1 Sept. What marketing can put behind that: the claims register so every number said on the phone matches the site, the routing so every region's enquiry lands with a person once you give one booking link per region, and the success stories your team already has in progress (Stove, ZIS, TVEO, ShiftMove) — unblocked the moment Tania/Fares/Eric hand over the missing data."
Pillar 3 — outbound content: "Your outbound works through process-pain conversations — the Landing Page Lab turns exactly that into pages: first offer built and audited, the next offer is days not weeks. The knowledge route (white paper, fireside chats, Trovarit/universities, filming — everything from the 7 Aug session with Heidi) feeds it."
The strategy point, straight: "You and Christoph have told me two different top stories — ERP-first versus four-areas-under-CFO-advisory. Both are defensible; the site can carry either; but sales and site must tell one story. Let's settle the candidates on the 19th and put the decision in front of Christoph on the 27th."
His open items from the 10th, read back: Cloudflare/GitLab rights + higher automation approval · the Jarrid escalation if needed · Heidi's ClickUp reprioritisation · November team-event photos · his review of the two team landing pages waiting since the 7th.

6All decisions to be made — numbered, with context; comment on any block

Strategy (the ones that outrank everything — S1 and S2 are above)

19 + 27 AUGS3 — Industry focus: adopt Lars's three (Software, Retail/E-Commerce, Light Manufacturing + software) or not? He named it himself as the missing piece of his strategy, with existing anchors (IT5 Solutions partnership → planned JV, 6PM/Teveo in retail). Christoph didn't address industries on the 13th. If adopted, it shapes every landing page and the install-base call scripts. Needs both their yes.

27 AUGS4 — The per-country product list, in writing. Christoph asked for exactly this himself (he corrected his own list mid-call: Everest/Aderis/Otheris/iplicit/Abacum all uncertain in the transcript). Nothing product-specific ships before this exists on paper. Owner: Christoph, in the room.

For Carsten, now

CARSTENC1 — v1.1 pick. Variant A never redraws real product screenshots (safe rule); variant B redrew the homepage collage (visibly more modern). Recommendation: A's rule as standard, B's hero as the exception. Compare.

CARSTENC2 — The two unsourced homepage efficiency percentages. No source exists; recommendation: off at cutover.

CARSTENC3 — Two gated hosts (admin interfaces; themes side-by-side live). Needs an Access app each so nothing public is minted — one go + one Cloudflare click (or the API token, C5).

CARSTENC4 — Team notes go out? Two English notes ready: the tech-stack explainer Abdel was promised on the 7th (his own open question — he expected WordPress to win "based on past precedent"; the note tells him it didn't, and invites him into Astro, which Lars explicitly endorsed on the 10th), and the AI-Specify status note. Plus the migration ruling: content survives, platform work doesn't.

CARSTENC5 — Security housekeeping: exposed review service token (needs your Cloudflare access to rotate) · preview-worker password redeploy (2 min, guards the client-seen cm./cmweb. flow) · wildcard *.elegant.work serving a stranger's WordPress on typos · the daily independent verifier stays paused until its environment gets the repository attached (~2 min of clicks).

For Lars (19 Aug, 11:00–13:00 — his own items read back + what only he can give)

LARSL1 — Cloudflare/GitLab rights + higher automation approval (his commitment, 10 Aug) — the go-live blocker.

LARSL2 — One booking link per region — turns the dead "any questions?" CTA into routed conversations; Christoph has asked for click-to-book repeatedly.

LARSL3 — Review the two team landing pages (NetSuite Partner DE, Salesforce — content-complete, waiting since 7 Aug) and the Lab's variant C.

LARSL4 — The six reference-customer names + the TVEO "we fixed it" case study (Eric owns the relationship; Tania/Fares/Eric owe Amr the success-story data since the 7th).

LARSL5 — Definitions + design polling. Sign the qualified-conversation/inbound definitions; and his own idea from the 10th — poll friends/family/customers on the design candidates ("ich würde gerne mit dem Design gehen, womit die Mehrheit geht"). We can set that test up within a day.

For Christoph (27 Aug, 11:00–16:30, Düsseldorf/Mönchengladbach)

CHRISTOPHCH1 — The demo-day block: the rebuild yes (logged) · S1 settled · S4 in writing · EBIT vs EBITDA for the €4m · faces confirmed · what may be said about the wider group · governance/rights model (presented as a proposal, not a question) · Jan's review scheduled with Christoph presenting · the filmed interview (location: Carsten owes it — garden ruled out, hotel lobby or the Gladbach house as fallback).