JUSTIN CHIA

Prompt
Engineering
Starter Kit

FULL 10-PAGE GUIDE

 The
Author

HELLO THERE! I'M JUSTIN.

I’VE BEEN CRAFTING PROMPTS FOR OVER A YEAR NOW.

PROMPT ENGINEERING HAS EVOLVED SO MUCH.

THIS PROMPT ENGINEERING STARTER GUIDE WAS CREATED TO KEEP

YOU UP TO DATE ON THE LATEST TACTICS.

IT WILL BE TARGETED TOWARD NEWBIES WHO ARE JUST STARTING

OUT.

BY THE END OF THIS EBOOK, YOU’LL LEARN MY BEST TECHNIQUES

AND BE A PROMPT ENGINEERING PRO!

LET’S GET STARTED!

MORE ON [Link]
 HOW TO START AN ONLINE BUSINESS

CHAPTER N.1

Prompt Anatomy:
The 4 Parts

You’ll learn what makes up a good prompt in

the 4 main parts of a prompt.
Prompt Anatomy: The 4
Parts
1/ CONTEXT
Set the scene! Provide relevant background information about the topic or situation.
This helps the LLM understand the purpose of the prompt and the world it's operating
in.
Example: "Imagine you are a customer service representative for a clothing store. A
customer has emailed you with a question about the return policy for a recently
purchased dress..."

2/ INSTRUCTIONS 3/ DESIRED TONE

Clearly define what you want the LLM Indicate the voice and style you want
to do. Be specific about the task at the LLM to adopt. Formal, informal,
hand and the format of the desired humorous, serious – the choice
output. depends on the context and
Example: "...Write a polite and audience.
informative email response explaining Example: "...Maintain a professional
the return policy and steps involved." and helpful tone throughout the
email."

4/ EXAMPLES
Provide additional guidance by including examples of successful outputs or similar
prompts that demonstrate your desired outcome. This helps the LLM grasp the
nuances of style and content.
Example: "You can reference our existing return policy webpage for details. Here's an
example of a similar email sent to a previous customer..."
pen_spark

CONTEXT INSTRUCTIONS DESIRED TONE EXAMPLES

 HOW TO START AN ONLINE BUSINESS

CHAPTER N.2

5 Starter Prompt
Templates

This section provides pre-built prompts to

quickly get you prompting.
HERE’S HOW IT WORKS
These prompt templates will guide you on how to craft a good prompt.

Simply follow the guiding instructions carefully to input details into a prompt.

Then enter them into LLMs and see your results!

N.1 Social Media Posts:

Platform: Specify the platform (e.g., Facebook, Twitter) and target audience.
Hook: Start with a captivating sentence or question to grab attention. (e.g., "Did you
know...?" or "Struggling with...?").
Content: Provide valuable and engaging content aligned with the platform's format (e.g.,
include visuals for Instagram or ask a thought-provoking question for Twitter).
Call to Action: Encourage interaction with a clear call to action (e.g., "Click the link to
learn more!" or "Share your thoughts in the comments below.").

N.2 Email Writing

Subject Line: Craft a catchy subject line that entices the recipient to open the email.
(e.g., "Free consultation: Boost your website traffic!").

Body: Write a clear, concise, and professional email tailored to the recipient and
purpose. (e.g., "Dear [Name], I'm reaching out today to discuss how our marketing
services can help you achieve your business goals. We specialize in [mention relevant
area] and have a proven track record of success. Would you be interested in a free
consultation to discuss your specific needs?").
N.3 Affiliate marketing
Topic: Define the blog post subject
and target audience.
Title: Craft a compelling title that
accurately reflects the content. (e.g., "
[Number] Proven Tips to [Desired
Outcome]").
Structure: Outline the key points you
want to cover in the blog post. (e.g.,
Introduction, Problem, Solution,
Benefits, Conclusion).
Style: Specify the desired tone (e.g.,
informative, humorous, inspirational).

N.4 Information Summarization

Source: Provide the text you want the LLM to summarize (e.g., a
news article URL or a lengthy research paper).
Length: Indicate the desired length of the summary (e.g., short
bullet points or a concise paragraph).
Focus: Specify the key points you want the LLM to prioritize in
the summary (e.g., main argument, research findings, important
events).
N.5 Script
Genre: Indicate the type of script (e.g., movie scene, play
dialogue).
Characters: Describe the key characters involved in the scene.
Plot Point: Outline the specific event or conversation you want
the script to focus on.
Tone: Specify the desired emotional tone (e.g., suspenseful,
comedic, dramatic).

IMPORTANT NOTE:
Remember: These are just a starting point!

Feel free to experiment and customize these templates to fit

your unique creative needs.

In the next section, I’ll share some professional-grade sample

prompts.
CHAPTER N.3

5 Professional-
grade Sample
Prompts

Lorem ipsum dolor sit amet, consectetur

adipiscing elit, sed do eiusmod tempor
incididunt ut labore et dolore magna aliqua.
5 Professional-grade
Sample Prompts
N.1 COLD EMAIL: PERSONALIZED OUTREACH WITH QUANTIFIABLE IMPACT

N.2 SOCIAL MEDIA POST: TWITTER WITH ENGAGEMENT OPTIMIZATION

N.3 BLOG POST: SEO-OPTIMIZED HEADLINE WITH AUDIENCE TARGETING

N.4 MOVIE SCENE: EMOTIONAL RESONANCE WITH CHARACTER ARCS

N.5 CUSTOMER CHURN PREDICTION: WITH ACTIONABLE INSIGHTS

How to Use These

Prompts?
As these are grab-and-go templates, you can simply copy them and paste them into
any LLM for immediate use.

YOUR LLM’S CHAT BOX

PASTE SAMPLE
PROMPTS HERE

ENTER A PROMPT HERE

 N.1 SALES (COLD EMAIL): PERSONALIZED
OUTREACH WITH QUANTIFIABLE IMPACT

Subject Line: Actionable Insights to Boost Your [Client's KPI] by [Percentage] (Data-
Driven)

Body:

Hi [Name],

I came across [Company Name] while researching industry leaders in [Client's Industry]. I
was particularly impressed by your recent achievement of [mention a specific
accomplishment].

At [Your Company], we specialize in helping businesses like yours achieve significant

improvements in [Your Value Proposition] through [Your Solution]. Our data shows that
past clients using our solution have experienced an average increase of [Specific
Percentage] in [Client's Key Performance Indicator (KPI)].

For instance, we recently partnered with [Similar Client Name] who faced challenges with
[Client's Pain Point]. By implementing our [Solution], they were able to [Quantifiable
Result] which directly translated to a [Specific Percentage] boost in [Client's KPI].

We understand that every business has unique needs. To see if we can be a good fit for
[Company Name], I'd love to schedule a quick 15-minute call to discuss your specific
goals and current strategies for improving [Client's KPI].

Call to Action:
Would you be available for a call next week? You can easily schedule a time that works
best for you through this link: [link to calendar tool].

Measurable KPIs:
Open Rate (% of recipients who open the email)
Reply Rate (% of recipients who respond)
Meeting Booking Rate (% of recipients who schedule a call)
Conversion Rate (% of leads who convert into paying customers)
N.2 MARKETING (SOCIAL MEDIA POST -
TWITTER WITH ENGAGEMENT OPTIMIZATION)

Content:
Did you know that [insert surprising statistic related to your industry]? [Link to Credible
Source] #IndustryHashtag #ValueProposition

Engagement Optimization:
Include a high-quality image or infographic related to the statistic.
Ask a thought-provoking question in the follow-up tweet to spark conversation. (e.g.,
"What are your biggest challenges related to [topic of the statistic]?").
Tag relevant industry influencers or publications to increase visibility.

Measurable KPIs:
Impressions (Number of times the tweet is seen)
Engagement Rate (Likes, Retweets, Clicks divided by Impressions)
Brand Mentions (Number of times your brand is mentioned in replies or other tweets)
Website Traffic (Click-through rate from the tweet to your website)
 N.3 BLOG POST (SEO-OPTIMIZED HEADLINE
WITH AUDIENCE TARGETING)

Target Audience: Small business owners in the e-commerce industry

N.3 BLOG POST (SEO-OPTIMIZED HEADLINE

Headline: The Ultimate Guide to Increasing Your Online Store's Conversion Rate by 20% in
30 Days (SEO-Optimized) [Year]
WITH AUDIENCE TARGETING)
Content:
Conduct keyword research to identify relevant search terms with high search volume
and low competition.
Structure the blog post using clear headings, subheadings, and bullet points for
improved readability.
Include high-quality visuals (screenshots, infographics) to break up text and enhance
understanding.
Optimize the content with relevant keywords throughout the post, but prioritize
providing valuable and informative content.
Conclude with a clear call to action, encouraging readers to download a free
resource, subscribe to your newsletter, or contact you for a consultation.

Measurable KPIs:
Organic Traffic (Number of visitors from search engines)
Time on Page (Average time users spend reading the blog post)
Bounce Rate (Percentage of visitors who leave after viewing only one page)
Lead Generation (Number of sign-ups for your newsletter or other offers)
 N.4 SCRIPT WRITING (MOVIE SCENE -
EMOTIONAL RESONANCE WITH CHARACTER
ARCS)

N.3 BLOG POST (SEO-OPTIMIZED HEADLINE

Scene: Two estranged best friends, Sarah and Michael, bump into each other after 10
years apart at a crowded coffee shop.
WITH AUDIENCE TARGETING)
Character Arcs:
Sarah: Has become a successful lawyer but feels unfulfilled and misses the close
bond she once had with Michael. (Yearning for connection)
Michael: Is now a struggling artist, haunted by past regrets and the breakdown of their
friendship. (Seeking redemption)

Emotional Resonance:
Capture the initial awkwardness and surprise of their unexpected encounter.
Gradually reveal the emotional depth of their characters through dialogue and
nonverbal cues (hesitation, longing glances).
Allow for a mix of humor (referencing old memories) and vulnerability as they begin to
address the reasons for their estrangement.

Measurable KPIs (indirectly through audience response):

Positive Reviews (Highlighting the emotional impact of the scene)
Audience Engagement (Tears, laughter, discussions about the characters' journeys)
 N.5 DATA ANALYSIS: CUSTOMER CHURN
PREDICTION WITH ACTIONABLE INSIGHTS

Goal: Identify key factors that contribute to customer churn for an e-commerce
platform and develop a machine learning model to predict churn with high accuracy.

Data Source:
N.3Customer
BLOGtransaction
POSTdata (SEO-OPTIMIZED HEADLINE
(order history, purchase amounts, frequency, etc.)
Customer profile data (demographics, location, preferences)
WITH AUDIENCE
Customer TARGETING)
support interaction data (tickets, complaints, resolutions)

Data Cleaning and Preprocessing:

Handle missing values using appropriate techniques (imputation, deletion).
Address outliers and data inconsistencies.
Encode categorical variables for machine learning compatibility.

Exploratory Data Analysis (EDA):

Analyze customer demographics and behavior patterns.
Calculate churn rate by product category, purchase frequency, and other relevant
variables.
Visualize data trends using charts and graphs to identify potential relationships
between churn and customer behavior.

Feature Engineering:
Create new features based on existing data (e.g., average order value, customer
lifetime value).
Reduce dimensionality if necessary to improve model performance.

Model Selection and Training:

Choose appropriate machine learning models for churn prediction (e.g., logistic
regression, decision trees, random forest).
Train and evaluate multiple models using cross-validation techniques.
Select the model with the highest accuracy, precision, and recall for churn
prediction.
CONT’D

Model Interpretation and Actionable Insights:

Utilize feature importance analysis to identify the most significant factors influencing
customer churn.
Segment customers into high-risk and low-risk churn categories based on model
predictions.
N.3 BLOG POST (SEO-OPTIMIZED HEADLINE
Develop targeted retention strategies based on identified risk factors (e.g., loyalty
programs for high-spending customers, personalized discounts for at-risk customers).
WITH AUDIENCE TARGETING)
Measurable KPIs:
Model Accuracy: Percentage of correctly predicted churn events.
Precision: Ratio of true positives (correctly identified churners) to total positive
predictions.
Recall: Ratio of true positives (correctly identified churners) to total actual churned
customers.
Reduction in Churn Rate: Percentage decrease in customer churn after
implementing targeted retention strategies based on model insights.

Additional Considerations:
Incorporate A/B testing to validate the effectiveness of proposed retention
strategies.
Regularly monitor model performance and retrain as needed to maintain accuracy
with evolving customer data.
 HOW TO START AN ONLINE BUSINESS

CHAPTER N.4

Fine-Tuning Tips:
Mastering the Art
of the Prompt

Learn the way of crafting the perfect prompt

for any use case.
 Fine-Tuning Tips: Mastering
the Art of the Prompt
Crafting a strong prompt is just the first step. To truly unleash the power of LLMs, you
need to fine-tune your prompts for specific needs. Here's how to adjust prompts for
maximum impact:

1. Keyword Magic: Refining Focus and

Relevance
Targeted Keywords: Carefully chosen keywords within your prompt act as guideposts
for the LLM.
Example: Want a persuasive blog post on the benefits of solar energy? Include
relevant keywords like "renewable energy," "cost savings," and "environmental
impact."

Keyword Placement: Strategic placement enhances the LLM's understanding.

Example: Instead of a generic "Write a blog post about solar energy," try "Write a blog
post informing readers about the environmental benefits of switching to solar energy."

Keyword Exclusion: Sometimes, excluding irrelevant keywords helps avoid unwanted

content.
Example: "Write a creative story about a robot, but avoid portraying it as a villain or a
threat to humanity."

2. Style Savvy: Shaping the Tone and

Voice
Descriptive Language: Describe the desired style using clear language.
Example: "Write a social media post in a humorous and lighthearted tone, perfect for
engaging our target audience."

Style Examples: Provide snippets or links to existing content that reflects the desired
style.
Example: "For reference, here's an example of a social media post with a similar
humorous tone: [link to example]."

Emotional Nuances: Convey the desired emotional tone for creative outputs.
Example: "Write a poem that evokes feelings of melancholy and nostalgia."
3. Tailoring the Output: Shaping Content
Length and Specificity
Desired Length: Indicate the preferred length of the output.
Example: "Generate a concise product description in 50 words or less."

Specificity through Details: Provide additional details to refine the output.

Example: Want a summary of a scientific paper? Include the paper's title, authors,
and specific aspects you want summarized (e.g., methodology, key findings).

Formatting Requests: Specify the desired format for the output.

Example: "Generate a bulleted list outlining the main points of the article."

Personal Note:
Remember that fine-tuning is an iterative process.

Feel free to experiment with different approaches and assess the LLM's output.

As you experiment, you’ll find better ways to craft powerful prompts over time.

By paying attention to keywords, style, and desired output details, you'll transform your
prompts from instructions into powerful tools for leveraging the full potential of LLMs.
 HOW TO START AN ONLINE BUSINESS

CHAPTER N.5

COMMON
PITFALLS:
PROMPTING
PERFECTION
Even the best don’t get these things perfect.
Take note of these tips!
Common Pitfalls: Prompting
Perfection
This table outlines common pitfalls to avoid when crafting prompts for Large
Language Models (LLMs):

Pitfall Description Potential Impact Solution

Provide a clear focus area, desired

LLM struggles to
outcome, and relevant background
Vague understand the
Overly Broad information. (e.g., "Write a blog post on 5
instructions or purpose and
Prompts travel destinations for budget-conscious
lack of context. generates irrelevant
millennials in 2024" instead of "Write
outputs.
something interesting about travel.")

LLM generates
outputs based on Double-check all facts and ensure the
Feeding the
Factual inaccurate information in your prompt is accurate
LLM incorrect
Errors information, leading and up-to-date. Use reliable sources for
details.
to misleading or data and avoid making assumptions.
nonsensical results.

Understand that prompt engineering is

Expecting a
an iterative process. Experiment with
magic formula Frustration and
Unrealistic different approaches and refine prompts
or instant disappointment with
Expectations based on the LLM's response. The LLM
expertise from generated outputs.
can't be an expert without your
the LLM.
guidance.

Overly Broad Prompts: Imagine asking a friend, "Tell me something about travel."
They might talk about their favorite vacation spot or the latest travel trend. For an
LLM, this lack of direction leads to confusion and irrelevant outputs.

Factual Errors: Think about giving your friend wrong directions. They might end up
lost or frustrated. Similarly, feeding the LLM inaccurate information leads to
misleading outputs.

Unrealistic Expectations: Don't expect the LLM to become an instant expert on a

topic. Even the best prompts require some back-and-forth refinement. Think of the
LLM as a powerful tool, but you're still the expert who provides the direction. By
working together, you can achieve great results.
 [Link]

Want more insights like this?

Join my newsletter:
The Juicer
You’ll receive weekly insights on AI, self-improvement, analytics, and productivity. No BS!

JOIN THE MASTERCLASS