Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Budget Planning — Ngan Sach Marketing · minhnv0807 | DecimalAI

minhnv0807/61-budget-planning

SkillScore50Maintained

Lap va quan ly ngan sach marketing: phan bo theo kenh / funnel / thang, ti le test vs scale, nguong tang - cat budget (nguong cat lo), quy du phong, budget tracker va review cadence. Dung khi nhac 'phan bo ngan sach', 'budget marketing', 'budget planning', 'chi bao nhieu cho ads'.

marketingby minhnv0807

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Marketing & Content pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/minhnv0807-61-budget-planning/SKILL.md

decimalai skills pull minhnv0807-61-budget-planning --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/minhnv0807-61-budget-planning/SKILL.md https://app.decimal.ai/s/minhnv0807-61-budget-planning/SKILL.md

Raw file: /s/minhnv0807-61-budget-planning/SKILL.md ↗ · machine-readable summary at /s/minhnv0807-61-budget-planning.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

22 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 25d ago

Yes — +36% lift · at higher cost+119% tokens0% turns50% of whole cases pass with skill (11/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-04	✗→✓	▲ Improved	157%	0%
case-06	✗→✓	▲ Improved	138%	0%
case-09	✗→✓	▲ Improved	127%	0%
case-10	✗→✓	▲ Improved	192%	0%
case-13	✗→✓	▲ Improved	153%	0%

Showing 5 of 22 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/10/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

minhnv0807 / ai-business-skills

skills/vi/61-budget-planning/SKILL.md

View on GitHub ↗

License MITImported 8/17/2026Version v1

Skill contents

SKILL.md

Budget Planning — Ngan Sach Marketing

> Budget khong phai con so co dinh — la cong cu ra quyet dinh: dat cuoc vao dau, rut ve khi nao. Leader phai biet: tong dang tieu bao nhieu, ROI tung kenh la gi, cai nao can tang, cai nao can cat. Goi 10-tinh-kpi-nguoc de tinh budget toi thieu tu revenue target truoc khi phan bo.

Thu thap thong tin

Doc .agents/product-marketing-context.md, ke hoach marketing va report ky truoc neu co. Neu thieu, hoi toi da 4 cau:

Budget tong va ky ke hoach? Tong bao nhieu, cho thang / quy / nam?
Revenue target ky nay? De tinh marketing % of revenue va budget toi thieu.
Kenh dang chay va ROI ky truoc? Meta / TikTok / Google / Zalo / Email / Agency — kenh nao ROAS/CPL tot nhat?
Co campaign lon nao can reserve budget khong? Launch, Tet, seasonal peak?

Nguyen tac

Budget follow performance — khong cung nhac theo plan ban dau. Kenh work → tang; kenh khong work → cat.
Luon giu 10-15% reserve — khong allocate het ngay dau ky. Reserve dung cho co hoi bat ngo hoac su co.
Moi dong chi ra phai co KPI do duoc — khong co "branding khong do duoc".
Khong phan bo deu — tap trung vao kenh work nhat; kenh moi chi cap budget test toi thieu (10-15%) truoc.
Tinh nguoc tu doanh thu, khong tu kenh — budget la ket qua cua bai toan revenue, khong phai con so tuy hung.

Quy trinh

Buoc 1 — Chon phuong phap thiet lap budget

| Phuong phap | Cach tinh | Uu diem | Nhuoc diem | |-------------|-----------|---------|------------| | % Doanh thu | Budget = X% doanh thu ky vong | Don gian | Khong phan anh giai doan tang truong | | Tinh nguoc tu muc tieu (khuyen dung) | Dat doanh thu → tinh nguoc CAC/CPL | Logic, co can cu | Can biet CAC/CPL lich su | | Theo doi thu | Benchmark ngan sach nganh | Tranh under-invest | Khong phan anh dieu kien rieng |

Cong thuc tinh nguoc 7 buoc (chi tiet dung 10-tinh-kpi-nguoc):

1. Doanh thu muc tieu
2. / AOV                      = So don can
3. / CVR chot (30-50%)        = So hot lead can
4. / CVR nurture (50-60%)     = So warm lead can
5. / CVR thu lead (20-40%)    = Tong reach can
6. Tong lead x CPL            = Budget ads toi thieu
7. + Content + Tool + Agency  = Total marketing budget

Lap 3 scenario: Conservative (CVR thap hon TB nganh, CPL cao hon) · Base case (bang TB nganh) · Optimistic (sau khi toi uu). Ke hoach chinh dung Base, chuan bi tam ly bang Conservative.

Buoc 2 — Phan bo tong quan theo bucket

| Bucket | % goi y | Muc dich | |--------|---------|----------| | Paid Ads (Performance) | 50-65% | Meta, TikTok, Google, Zalo | | Content Production | 10-15% | Video, design, copywriting | | KOL / KOC / Seeding | 5-10% | Social proof, UGC | | Agency / Freelancer | Theo scope | Neu co | | Tools / Software | 2-5% | Ads tools, design tools, email/CRM | | Campaign Reserve (du phong) | 10-15% | Co hoi bat ngo / su co — KHONG allocate truoc |

Buoc 3 — Phan bo theo kenh

Benchmark SME Vietnam (dieu chinh theo nganh va data rieng):

| Kenh | % ngan sach ads | Objective | Expected CPL/ROAS | |------|-----------------|-----------|-------------------| | Meta Ads (FB + IG) | 50-60% | Lead / Purchase | | | TikTok Ads | 15-20% | Awareness / Lead | | | Google Ads / SEO | 10-15% | Search intent | | | Zalo Ads / OA | 5-10% | Retarget / Local / Nurture | | | Kenh moi (test) | 10-15% | Validate | |

Logic phan bo: kenh dang co ROAS/CPL tot → tang ti trong; kenh moi → budget test toi thieu truoc; khong phan bo deu.

Buoc 4 — Phan bo theo funnel va ti le test vs scale

Budget split chuan performance trong tung kenh:

| Muc dich | % | Ghi chu | |----------|---|---------| | Testing (creative + audience moi) | 30% | Luon co dong test — creative fatigue den nhanh | | Scale (winner da xac nhan) | 50% | Chi scale campaign co data thang | | Retargeting | 15% | Warm audience — CPL re nhat funnel | | Lookalike | 5% | Mo rong tu tep converter |

Theo funnel: TOFU (reach/awareness) — MOFU (lead/nurture) — BOFU (conversion/retarget) — ti le tuy giai doan: launch nghieng TOFU, mature nghieng BOFU + retention.

Buoc 5 — Phan bo theo campaign va thang

| Campaign | Thang | Budget | Objective | KPI | |----------|-------|--------|-----------|-----| | Always-on (retarget + nurture) | Ca ky | | | | | Campaign A (launch/seasonal) | | | | | | Testing budget | Ca ky | | Thu kenh / angle moi | | | Reserve | — | 10-15% | | |

Buoc 6 — Nguong tang / cat budget (decision rules)

Tang budget khi (du CA 2 dieu kien):

ROAS > target x 1.5 trong >= 7 ngay lien tiep, HOAC CPL < target x 0.7 va volume du
Audience chua bao hoa (frequency < 2.5)
Muc tang: +20-30% moi lan, khong gap doi dot ngot

Cat / pause budget khi (nguong cat lo):

| Dieu kien | Hanh dong | |-----------|-----------| | ROAS < break-even sau 14 ngay | Cat kenh/campaign, chuyen budget sang kenh work | | CPL > target x 2 sau 7 ngay test | Pause, thay creative/audience truoc khi bat lai | | Frequency > 4 | Audience saturation — dung scale, refresh creative | | Spend > 110% ke hoach thang | Cap daily budget, review lai phan bo |

Vi du CPL rule hang tuan (nganh dich vu local — can chinh theo AOV cua ban): CPL < 12K & freq < 2.0 → WIN, scale +20-30% · 12-15K → acceptable, theo doi · 15-18K & freq 2.0-2.5 → optimize, thay creative · > 18K & freq > 2.5 → PAUSE, replace.

Khong bao gio tang budget khi CPL dang xau — fix creative/tracking truoc.

Buoc 7 — Tracker va review cadence

Budget tracker theo thang:

| Thang | Budget plan | Thuc chi | Variance | ROAS/ROI | Ghi chu | |-------|------------|----------|----------|----------|---------| | Thang 1 | | | | | | | Thang 2 | | | | | | | Thang 3 | | | | | | | Tong quy | | | | | |

Review cadence:

| Nhip | Viec | Output | |------|------|--------| | Hang tuan | Xem CPL/ROAS tung kenh — ap decision rules Buoc 6 | Dieu chinh ngay neu cham nguong | | Hang thang | Review tong phan bo, budget reconciliation plan vs actual | Rebalance cho thang sau | | Hang quy | Review chien luoc phan bo tong the, LTV:CAC | Dieu chinh strategy |

Suc khoe don vi kinh te — LTV:CAC:

| Ti le | Y nghia | Hanh dong | |-------|---------|-----------| | < 1:1 | Dang lo tren moi khach | Dung scale, fix product/price | | 1:1 → 3:1 | Hoa von den kha thi | Toi uu conversion, tang AOV | | > 3:1 | Healthy | Bat dau scale | | > 5:1 | Rat tot | Tang ngan sach ads manh |

Cau truc ket qua

Ten file: budget-plan-[ten-thuong-hieu]-[YYYYMMDD].md — gom: I. Budget overview (tong, revenue target, marketing % of revenue, ky) · II. Phan bo bucket · III. Phan bo kenh · IV. Test vs scale split · V. Phan bo campaign/thang · VI. Decision rules + nguong cat lo · VII. Tracker · VIII. Review cadence.

Lien ket skill

10-tinh-kpi-nguoc: tinh budget toi thieu tu revenue target — chay truoc buoc phan bo.
00-ke-hoach-mkt: budget plan la Phan 4 chi tiet hoa cua ke hoach tong.
54-media-plan: trien khai budget ads thanh media plan cu the theo campaign/adset.
55-scaling-ads: quy trinh scale khi cham nguong tang budget.
21-audit-ads-performance: audit khi kenh cham nguong cat lo lien tuc.
07-bao-cao-marketing: budget reconciliation trong report thang.

Checklist chat luong

] Budget tinh nguoc tu revenue target, co 3 scenario
] Co du 3 lop phan bo: bucket → kenh → campaign/thang
] Ti le test vs scale ro rang (Testing 30 / Scale 50 / Retarget 15 / LAL 5)
] Reserve 10-15% duoc giu lai, khong allocate truoc
] Nguong tang va nguong cat lo co con so cu the (ROAS, CPL, frequency, so ngay)
] Kenh moi chi nhan budget test 10-15%
] Moi dong budget co KPI do duoc di kem
] Co tracker plan vs actual theo thang
] Review cadence tuan/thang/quy co output ro
] LTV:CAC duoc kiem tra truoc khi quyet dinh scale

Illustrative Want to see what minhnv0807/61-budget-planning would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

9/2/2026, 23:19 UTC · 22 cases

Lift

+36%

50% vs 14%

Agent turns

1.0 vs 1.0

Tokens

+119%

7,089.09 vs 3,242

fail → pass · 8pass → pass · 3fail → fail · 11

Per-case results (22)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-20 An online bookstore, PageTurner, wants to set up automated email marketing workflows in Klaviyo for abandoned cart recovery and customer onboarding. Please design the trigger logic, delay timing, and email sequence content.	fail→fail	27,511	26,771	-3%	1	1	0%	3,317	6,891	+108%	0	0	—
case-01 Tôi là Marketing Manager của một thương hiệu thời trang local brand. Chúng tôi đang chuẩn bị lập kế hoạch ngân sách marketing cho Quý 3 với mục tiêu doanh thu 3 tỷ VND. Dữ liệu kỳ trước cho thấy Meta Ads và TikTok Ads đang mang lại CPL tốt nhất, đồng thời sắp tới chúng tôi có một chiến dịch ra mắt bộ sưu tập mới. Hãy giúp tôi xây dựng bản phân bổ ngân sách hoàn chỉnh dạng markdown. Bản kế hoạch cần tính toán ngược từ mục tiêu doanh thu với các kịch bản khác nhau, phân chia ngân sách theo các nhóm chi phí, phân bổ cụ thể theo từng kênh quảng cáo, đưa ra tỷ lệ giữa thử nghiệm và mở rộng, đặt ra bộ quy tắc cùng ngưỡng chỉ số cụ thể để tăng hoặc cắt giảm chi phí quảng cáo, đi kèm bảng theo dõi thực chi hàng tháng và lịch kiểm tra định kỳ.	fail→fail	34,241	85,593	+150%	1	1	0%	5,309	10,777	+103%	0	0	—
case-02 Tôi chuẩn bị duyệt ngân sách marketing cho doanh nghiệp cung cấp giải pháp B2B của mình trong quý tới với mục tiêu doanh thu kỳ vọng là 1.5 tỷ VND. Vui lòng lập cho tôi một tài liệu kế hoạch ngân sách chi tiết. Tài liệu cần bao gồm: phân tích mục tiêu doanh số để tính ra chi phí tối thiểu, chia phân khúc ngân sách cho các nhóm công việc chính và dự phòng, phân bổ chi tiết theo từng kênh media, thiết lập tỷ lệ ngân sách cho các giai đoạn phễu tiếp thị, xây dựng các tiêu chí kỹ thuật cụ thể để quyết định khi nào cần dồn thêm tiền hoặc dừng chiến dịch, cuối cùng là bảng tracker đối soát thực tế và nhịp độ review cho đội ngũ.	fail→fail	40,508	31,357	-23%	1	1	0%	5,692	8,043	+41%	0	0	—
case-08 A mobile accessory retailer, TechGear, is setting up paid ads budget distribution across ad sets. The media buyer wants to use an 80/20 split between winning ads and testing ads. Provide the recommended percentage split across testing, scaling, retargeting, and lookalike campaigns.	fail→fail	17,961	23,967	+33%	1	1	0%	2,091	5,585	+167%	0	0	—
case-03 Thương hiệu mỹ phẩm DTC của chúng tôi cần một bản hoạch định chi tiêu marketing cho 6 tháng cuối năm dựa trên target doanh số được giao. Hãy lập giúp tôi tài liệu kế hoạch ngân sách hoàn chỉnh theo cấu trúc chuẩn. Yêu cầu đầu ra phải có: tổng quan ngân sách đi kèm các kịch bản doanh thu, tỷ lệ phân bổ chi phí theo các nhóm hạng mục và kênh chạy ads, nguyên tắc chia dòng tiền giữa chiến dịch test và scale, các quy tắc hành động kèm ngưỡng con số rõ ràng cho việc tăng/cắt ngân sách theo diễn biến thực tế, cùng mẫu bảng theo dõi chênh lệch kế hoạch vs thực chi và quy trình đánh giá chỉ số sức khỏe tài chính.	fail→fail	44,343	46,807	+6%	1	1	0%	6,909	10,738	+55%	0	0	—
case-04 An e-commerce furniture brand, HomeDeco VN, is planning a $20,000 monthly marketing budget. The team wants to distribute 100% of the funds across active ad channels immediately at the start of the month to maximize reach. Please draft their general bucket allocation structure in markdown, detailing how reserve funds should be handled upfront.	fail→pass	18,970	30,942	+63%	1	1	0%	2,939	7,563	+157%	0	0	—
case-05 An online education platform, Edutech Edu, wants to launch on a new ad channel (LinkedIn Ads) and is considering dedicating 40% of their ad spend to test it alongside existing channels. Produce the channel allocation recommendation for Edutech Edu.	pass→pass	28,316	35,070	+24%	1	1	0%	3,618	7,413	+105%	0	0	—
case-06 A SaaS brand, CloudDesk, wants to know when to scale campaign budgets. The ads lead suggests doubling daily budget whenever ROAS stays above target for 2 days. Draft the exact decision rules for scaling ad spend, including performance ratio, required consecutive duration, frequency threshold, and safe spend increase percentage.	fail→pass	12,821	22,050	+72%	1	1	0%	2,517	5,989	+138%	0	0	—
case-07 An online shoe brand, SneakerX, wants clear criteria on when to pause underperforming campaigns. The marketing executive proposes pausing campaigns after 2 days if CPL increases by 20%. Formulate the exact cut/pause thresholds for CPL inflation, audience saturation frequency, and ROAS breakeven windows.	fail→fail	27,198	29,509	+8%	1	1	0%	3,488	6,965	+100%	0	0	—
case-09 A healthcare supplement company, NutriLife, targets 2 billion VND in revenue for Q4. Provide the step-by-step mathematical logic to derive the required minimum marketing budget from this revenue goal.	fail→pass	23,297	28,258	+21%	1	1	0%	3,456	7,858	+127%	0	0	—
case-10 A premium coffee subscription brand, RoastMaster, is forecasting Q1 marketing expenses. The CFO insists on using a single baseline calculation. Produce the scenario planning structure required for budget setup.	fail→pass	23,335	37,501	+61%	1	1	0%	2,926	8,530	+192%	0	0	—
case-11 A skincare local brand, GlowSkin, is dividing its paid media budget across channels. The team wants to allocate 50% to TikTok Ads and 20% to Meta Ads. Provide the standard paid media allocation percentages for Vietnamese SME performance marketing.	fail→fail	14,658	25,019	+71%	1	1	0%	2,355	5,743	+144%	0	0	—
case-12 A subscription box startup, SnackBox, wants to decide whether to aggressively increase marketing budget. Their current LTV:CAC ratio is 2:1. Provide the evaluation framework for unit economics and state what LTV:CAC ratio is required before aggressive scaling.	pass→pass	18,545	20,170	+9%	1	1	0%	2,553	5,759	+126%	0	0	—
case-13 A luxury watch retailer, ChronoVN, wants to organize their marketing budget oversight schedule. Detail the review cadence and specific deliverables required weekly, monthly, and quarterly.	fail→pass	20,202	25,453	+26%	1	1	0%	2,468	6,237	+153%	0	0	—
case-14 A fitness apparel brand, FitFlex, needs their finalized quarterly budget document. Show the header metadata and specify the standard file naming convention for saving this document.	fail→fail	7,378	42,006	+469%	1	1	0%	1,237	9,474	+666%	0	0	—
case-15 A food delivery app, QuickBite, is preparing its budget document for investors. Outline the required main sections (I through VIII) in the standard budget document structure.	fail→pass	18,572	29,936	+61%	1	1	0%	2,138	8,552	+300%	0	0	—
case-16 An organic juice brand, FreshJuice, is setting up its total marketing expenditure categories. The team proposes putting 80% into paid ads and 2% into content production. Provide the recommended standard bucket percentages across Paid Ads, Content Production, KOL/KOC, Tools, and Reserve.	fail→pass	18,267	20,442	+12%	1	1	0%	2,189	5,145	+135%	0	0	—
case-17 A handmade leather goods shop, CraftLeather, has been running Google and Meta ads for 2 weeks. One campaign has ROAS below breakeven. Specify the decision rule and time window for cutting or pausing campaigns operating below breakeven ROAS.	pass→pass	14,692	19,086	+30%	1	1	0%	2,261	5,494	+143%	0	0	—
case-18 A kids toy brand, ToyWorld, notices one of their monthly campaigns has spent 115% of its planned monthly budget. State the decision rule when spend exceeds monthly budget allocations.	fail→pass	13,985	13,621	-3%	1	1	0%	1,327	3,964	+199%	0	0	—
case-19 A fashion brand, UrbanThread, wants compelling Vietnamese ad copy headlines and primary text scripts for their upcoming summer sale campaign on Meta Ads. Please write 3 ad copy variations with hooks and calls to action.	fail→fail	64,227	33,135	-48%	1	1	0%	2,361	5,272	+123%	0	0	—
case-21 An electronics reseller, TechMart, wants a technical SEO audit of their website architecture, focusing on XML sitemaps, canonical tags, and crawl errors. Provide a technical SEO audit report with fixes.	fail→fail	36,269	49,886	+38%	1	1	0%	5,729	7,459	+30%	0	0	—
case-22 A tea franchise, LeafTea, needs graphic design prompt concepts and visual layout templates for their Instagram carousel posts. Produce visual design specs and prompt instructions for visual assets.	fail→fail	33,383	30,527	-9%	1	1	0%	4,444	6,509	+46%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +36 percentage points is the difference between those two pass rates over the 22 comparable cases.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 13.6%
Pass rate with the skill: 50%
Cases attempted: 22
Cases counted toward lift: 22
Cases that regressed: 0
Skill version measured: v1
Run completed: 9/2/2026, 23:19 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-04

Prompt

An e-commerce furniture brand, HomeDeco VN, is planning a $20,000 monthly marketing budget. The team wants to distribute 100% of the funds across active ad channels immediately at the start of the month to maximize reach. Please draft their general bucket allocation structure in markdown, detailing how reserve funds should be handled upfront.

Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.

Show both answers

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run50

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns+121% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 25 days ago

Other measured skills in the registry, with their headline benchmark lift.