Follow these 12 steps roughly.

1) Clarify functionality of search engine
~ Ask questions about what specific are concerns of children under 14 yrs. Is it pornography, etc?
~ Ask if it should be an app or browser page that is the primary use case.
~ Ask where it would be launched: country
~ Do parents want full control of what content is shown to kids & hours.

2) Clarify exact segment & question
~ Ask which age group we should target 1 to 3, 2 to 5, 5 to 8, 8 to 13, etc. This is important so that the content can be tailored accordingly
~ Also, this impacts the exact interface. For kids around age 3, the interface would pretty much be a touch interface with pictures. For kids around 8, it could be a mix. Also, for kids around 12 to 14, we could assume they would have a mobile phone perhaps.
~ Are kids using this search engine for entertainment or for education or both.
~ Do parents want full control of what content is shown to kids & hours?

3) Define target user Persona
~ Jim from Austin, TX
~ Kid age 2.5
~ Lives at home with parents
~ Gone to PRESCHOOL most of the day
~ Has access to iPad
~ Interests: Cartoons, Numbers, Alphabets, Cars

4) Map user journey
~ Depending on the age group chosen. Say 2 to 5, we should map the journey for an IPad based mobile search engine app.
~ User opens iPad for image based search app
~ User opens app and is immediately presented with an option to browse specific categories of content: Science, Math, Cartoons, etc.
~ User browses content via images & arrives at the content they want to see

5) Identify Metrics that should matter
~ Number of times children use app to search relevant content.
~ D/M/Y usage in terms of:
>>> Tapped icons
>>> Viewed/read content
>>> Rating in app store
>>> User feedback in app store

6) Identify use cases & pain points
~ Kids want clickable categories. Current solutions don’t do that.
~ Parents want pornography, stalkers & other bad elements removed like chatrooms. Current platforms are geared for people of all age.
~ Parents should be able to control the hours. Current solutions are not geared towards parents’ controls.

7) Identify solutions
~ Create category based image search engine so that kids can navigate by clicking
~ Create parents dashboard so that kids’ access can be controlled
~ Ensure filters can be applied in terms of what content is available in this search engine
~ Recommendation engine based on past browsing history

8) Prioritize Solutions
Prioritize based on cost, ROI & time to build after MVP has been defined.

The baseline MVP would have to to build the content & categories first so that user feedback can be collected about that functionality. So we would prioritize like this:
1~ Create category based image search engine so that kids can navigate by clicking
2~ Ensure filters can be applied in terms of what content is available in this search engine

The remaining solutions would fall beyond the MVP. Therefore these would be evaluated against ROI, Cost & time to build.

3~ The parents dashboard : Low cost, ROI high say 10%, time to build = 2 months

4~ Recommendation engine – High cost, ROI high say 10%, time to build = 6 months

9) Validate
~ Design some wireframes, prototypes share with target children segment
~ Get feedback to improve designs

10) Iterate
~ Build updated designs
~ Arrive at final design
~ Build actual solution after design has been throughly vetted via wireframes, mockups, etc.

11) Launch & Measure
~ Launch to smaller audience
~ Validate their experience in production
~ Start tracking key metrics & get their early feedback
~ Incrementally launch to a larger audience OR start building the 2nd half of the solution for parents dashboard and then incrementally launch.

12) Do post-launch followup to keep an eye on key metrics
~ Keep an eye on metrics over time.

Going to assume that this search engine will be for all types (websites, videos, images..). Also going to make the assumption that users will use this on mobile and desktop.

Users:
1) Children looking to get help/do research for homework
2) Younger children looking to find videos of their favorite cartoons / superheroes
3) Children looking to find music videos of their favorite bands
4) parents

for the purposes of this answer, I’ll assume the target customers are the children trying to do research for their homework.

Use-cases:

1) Search the web to learn more about a particular topic
2) Search for the answer to a particular academic question
3) Search the web to get tutorials on a particular topic
4) Search the web for images and videos that they can use for their presentation
5) Get help in knowing which is the best source of truth for an answer
6) Get help in deciding which sources can be cited on a paper
7) Get answers instantly, and from anywhere

Painpoint: Overwhelmed with the amount of information provided

Prioritization by customer benefit and complexity:
highest: search for the answer to a particular academic question, get answers instantly from anywhere

Solutions:

1) Search engine app / site that provides answers to academic questions directly by tapping into knowledge graph (like Wolfram Alpha, but for a wider variety of topics).
Pros: children save time to get answer, can click links to learn more if they need to
Cons: Huge upfront cost to integrate wolfram alpha like capabilities

2) Search engine app/site that prioritizes searching in websites that are known to be relevant to children’s education – Quora, educational blogs..
Pros: Links and mentions in “recommend” websites get higher weight, which means more relevant material is surfaced;
Cons: redoing of the search algorithm

3) Search engine app / site that has no ads – perhaps subscription supported to NOT influence anything but the top result, get the resourcing to optimize results for children, hide inappropriate ads / content
Pros: Keep children focused on their search and hide explicit/inappropriate content
Cons: Not sure if this would work from a business viability standpoint when users are used to search engines being free

Prioritization by customer benefit and cost: #1 (search engine that provides more answers upfront to academic questions)

To further understand the I would want to ask few clarifying questions
1. When we say web search engine, are we referring to something like Google search/Bing? I will assume the answer to be yes.
2. On what platform are we going to build the search engine? Mobile/Desktop/Tablet? I will park this question for now and will address while understanding our target group.
3. We have mentioned that search engine should be for children below 14 years of age. Do we have any specific range, or we will be addressing the requirements of all age groups below 14? For now, I will assume 10 – 14 years.

Now I would like to understand what problem are we trying to solve, what will differentiate our product with that of existing products in market? I will assume that current product is not that user friendly for our target group and does not specifically cater to their needs.

Now I would like to know for whom we are building the product for:

1. Children in age group of 10 -14 years, I will assume they can use existing product (Google/Bing search engine) in the market. The understanding and knowledge level of will vary, I will assume our target group is a regular school going kid who understands basic level of English.
2. Parents who would use the search engine to find information related to kid’s education, their career
3. Teachers who would like to use search engine to teach kids and may be also give reference from search results
Now I will try to address my previous question of platform, since smartphone penetration has increased tremendously I will design the web search for Mobile.
After I have understood who our users will be, I will try to understand their behavior and needs in depth
1. Children in age group of 10 -14 years
a. Current product in the market assumes that person has basic comprehending capability, but children in age of 10-14 years are still in basic schooling phase. Hence, they lack ability to go through several results and decide which ones are the most suitable.
b. They might not clearly articulate what they are looking for
c. Short attention/focus duration
d. Visual content is more appealing
e. Susceptible to hacking
2. Parents
a. Would not want their kids to waste too much time on it
b. Want to ensure that there is no malicious content displayed
c. If possible, monitor the activity of kid on search engine
3. Teachers
a. Would want to recommend few links to their students/parents
b. Would want to search for latest development on the topic going on in class

Now that we have understood the needs to our users, I would like to list down our product features, for this I will try to come up with basic features based on above needs and prioritize the features

Basic Features
1. Use profile (age, country, city, school, interests) of user to return more relevant results
2. Recommendations from teachers and friends shown first
3. Video, image recommendations on top
4. An AI mechanism to understand the intent of the search instead of matching the words for searching
5. When hovering mouse over the link summary of the article to be explained – It can be either in form of text or voice
6. A feature to have one permanent and one dynamic/adhoc password
7. A dashboard showing the category of searches, time spent on searches
8. This can be further enhanced to see what other kids in the same age group are searching which can help parents/teachers/(kids themselves) in assessing their kids and taking any action
9. A separate tab listing down latest development/history/Basic information/Intermediate/advanced level

Prioritizing the features

1. Basic feature
2. Differentiating
3. Easily achievable
4. Complex/High cost
5. Complex/High cost
6. Easy, Low cost
7. Easy, Low cost
8. Medium complexity, low cost, high return can be made a paid feature as well
9. Medium complexity, low cost, high return can be made a paid feature as well
Hence,
Priority 1 – Features – 1,2 ,3 ,6,7,8,9
Priority 1 – Features – 4
Priority 1 – Features – 5

Summary
We will build a mobile search engine for children in the age group of 10-14 years which will allow use demographics, recommendations from teachers to show video and image results on top. A dashboard will be generated showing frequent searches and generated useful insights to identify areas of improvement. In addition, there will be tabs which would further classify the results retrieved into basic/intermediate/advanced/latest development.

To clarify this Google product design question, we want to build a search engine for children below the age of 14. I have a few questions that might affect the experience of this search engine:

Is this search engine general purpose, or geared towards jobs, or hobbies?
What country are we building this search engine for?

I’ll assume we’re building a general purpose search engine with the goal of providing children 14 and under the same capabilities of a search engine that a person from any other age group might use but tailored to the needs of this specific age group. I’ll also assume these users know how to use a computer since they wouldn’t need to use a search engine if they didn’t know how to use a computer.

Now, I’d like to identify and discuss the users that might be using this search engine.
1. Children 14 and under
This is the main group we’d be building the engine for. Children in this age range have
-varying degrees of internet awareness
-different education levels (reading vs writing) and typing ability
-are mostly interested in games and social activities
-are vulnerable to attacks by hackers

2. Parents
-Usually when children 14 and under have access to a computer, their parents have given them this access
-This user group cares about the safety of their children
-They don’t spend as much time doing the same activities on the search engine as their children do

I think we should focus on the children users for now. With that being said, I’d like to report on some of those use cases.

Children under 14
1. Search for and discover content
2. View and understand results for the content they search for so they can quickly get to what they’re looking for
3. Will want to know what their friends are up to so they’ll have shared experiences

Not that I’ve reported on some requirements, I’d like to discuss some possible solutions.

Children under 14
User Experience
1.
a) Word cloud search: this would show words, and concepts that other children are searching for and allow kids to click on them. This way kids can a) discover popular content and be in the know b) easily click to initiate a search without needing to type in anything
b) Topic search: this is similar to the first idea, but instead of individual words, kids would only see curated topics like (sports games, puzzle games, disney channel stars) and they’ll be able to go a few levels deep to get more specific for each category. This takes the guess work out and allows children to get to content that they care about quickly.
c) Visual auto complete: this would allow children to see actual images of the things that they are typing in once the fetcher gets to a certain level of confidence. When this happens the children can click on an item, and the search engine will complete the search. This will help make children aware of what they are doing and make it easier to execute a search.

2.
a) Visual results: this would ditch the common way we show results today (links and text) in favor of showing previews of the actual websites. This allows children to quickly identify what site they want to visit without needing to read lengthy descriptions.
b) Smart assistant in results: this would be a quick snippet of text for every returned result that would give the child a high level explanation of what they can do on the site before they click it

3.
a) Connect social: this would allow children to connect their social accounts of choice and give them the ability to see sites that their friends have visited on the default search page, but would require opting into the ability.

Search Engine Tech
4.
a)Crawler filter: this search engine should only be crawling sites that are age appropriate. One way this can be done is from the information itself, another way is from looking at the demographics of the users that go to the site itself. It’ll need to ignore any domain that doesn’t fit the children under 14 demographic or that has content that is inappropriate or may be abusive

I think that’s a good list of features. Let’s spend some time prioritizing them based on impact to cost ratio.

1a) Quick Win: High Impact, Low Cost. It’s fairly easy to keep a counter for queries per word, and showing these queries up front would provide a lot value for children.
1b) Must Have: High Impact, High Cost: Relative to the item before this feature would require some more processing to our data set to be able to surface up the right content, and provide further levels of granularity.
1c) Money Pit: Low Impact, High Cost. I think since children are already typing and have somewhat of an idea this won’t provide as much value, and it may be expensive to not only get the images, but be able to surface the right image for each word.

2a) Quick Win: High Impact, Low Cost. I think since most websites are cached it’d be pretty easy to show website previews for what users have searched for, and to make this the primary showing for the results would also be easy put together but make it really easy for children to consume.

3a) Money Pit: Low Impact, High Cost. I think showing a users friends activities could be compelling but we can assume that when we generalize behavior for an age group the behavior is similar so this may not add too much value and be costly.

4a) Must Have: High Impact, High Cost. This is absolutely necessary to power the experience users will see. The technology exists.

Some tradeoffs with these features include:
-not all children can see or are visual so we’d need accessibility affordances in the future
-the curated categories may not be what children want

To summarize, we want to build a search engine for children 14 and under with the goal of making an experience that meets their needs, and allows them to quickly and easily find content. We can do this by including the features 1a, 2a, 1b, and 4a in our MVP.