Answering Problem Solving Behavioral Questions
Learn how to answer "Tell me about a time..." styled questions for evaluating your problem solving skills.
As mentioned in our behavioral interview preparation overview, problem solving and drive for results is one of the 8 main categories of questions to prepare for.
In this guide, you will learn how to tackle them:
- Evaluation criteria in detail
- Abstracting possible questions into common themes
- Suggested answer framework
- Possible nature of follow-up questions
- Sample questions and answers
Evaluation criteria in detail
Problem solving and drive for results are grouped together into 1 category as they are highly related in nature. When describing one's experience in solving problems or to achieve certain results / objectives, their mindset or drive to do so can also be inferred.
When rating candidates under this category, interviewers are often looking at the following criteria:
- Identifying best solutions and executing on them
- Identifying the right problems
- Identifying the most critical objectives
- Resourcefulness and data-driven mindset
- Creativity and innovation
- Identifying tradeoffs and sustainable solutions
- Measuring results, iterating and following through
- Impact-driven mindset
- Proactivity to make progress despite obstacles or roadblocks
- Influencing others to deliver against objectives
- Balancing analysis with decisive action
Abstracting possible questions into common themes
As mentioned in our behavioral interview preparation overview, it is impractical to prepare answers specifically for every behavioral question out there. However, by batching specific questions into similar themes and preparing stories that cover a large number of question requirements, we can reduce the number of stories to prepare to around 3-5 stories.
There's simply too many types of problem solving and drive for results behavioral questions that could be asked, such as:
- Can you tell me about a time when you had to use data to drive engineering decisions?
- Can you provide an example of a time when you had to troubleshoot and fix a complex issue in a project?
- Can you describe a time you creatively solved an engineering problem or achieved meaningful metric improvement?
- Can you describe a time when you had to make an important engineering decision and how you decided between tradeoffs?
However, when we look at 80% of questions under this category, questions typically ask for how a problem was solved or a specific trait required for effective problem solving, such as creativity, using data or trade-off evaluation. The source of the problem or objective, as well as whether the candidate was resilient in the face of obstacles can be inferred from answers to these questions as well.
Suggested answer framework
As always, the STAR format is the simplest and most effective framework that we recommend to structure your story.
Despite the large number of question possibilities, we can abstract all the requirements for problem solving behavioral questions by thinking about a robust problem solving process:
- Problem identification: Identifying the correct root cause or root problem to focus on from surface level symptoms
- Metrics/target setting: Identify key metrics that would signify success in solving the problem (if any)
- Information gathering: Resourcefulness in gathering information from different sources and a data-driven mindset (using data to drive decisions)
- Solution brainstorming: Creatively generate solutions that solve the root cause
- Solution evaluation: Evaluate trade-offs of each solution and select the most optimal one
- Monitoring and adjustment: Monitor effectiveness of the solution by measuring key metrics. Adjust strategy if needed
Hence, you just have to ensure that at least 1 of your prepared stories or projects covers all of the above steps. In doing so, the story could be reused for all affiliated traits under problem solving, such as data-driven mindset, resourcefulness, creativity, and working with feedback from users.
Of course, you might need to tailor the details of your story to focus on the specific aspect asked in the specific question.
We recommend you to also pick problem solving stories that can doubly serve to get signals on the following:
- Proactivity/initiative-taking: You had taken the initiative to look into the problem, gather information and solve it
- Leadership: You had led the problem solving process
- Teamwork: You had to work as part of a team to solve the problem
Sample story
Here's a sample story using the STAR format.
Situation: I was the tech lead for an e-commerce website selling luxury goods. The website was built as an Angular 1.5 single-page application. In recent years, the product was showing its age – developer experience was not great and the website performance was poor. Initial loading speed was over 3 seconds and the conversion rate was around 0.8%.
Task: I was tasked with improving the performance and conversion of the website.
Actions
- Problem identification
- Conversion is tied to website performance and UX
- Website performance has been on a gradual decline over the past few years as there wasn't a person to oversee the overall architecture of the website and track performance metrics
- UX hasn't been looked at in a while
- Information gathering
- Looked at nature of bugs in the past year, categorized them according to their root causes to identify hotspots and major problematic areas.
- Gathered feedback from team regarding areas of improvements.
- Together with the product manager, ran a brainstorming session with the team to think of ways to improve.
- In order to improve, firstly we need to know how we are doing:
- Double checked that our performance and conversion tracking was working correctly.
- Started tracking new metrics from Lighthouse and Core Web Vitals.
- Worked with Data Scientists to come up with dashboards for performance and conversion and gained some insights:
- Identified that some countries had lower conversion rates.
- Mobile users had lower conversion rates as compared to desktop.
- Worked with UX Designers and UX Researchers to identify problems in the end-to-end shopping experience on the website.
- UI elements were too spaced apart and required a lot of scrolling which affected bounce rate because some users didn't bother scrolling.
- Solution brainstorming
- User interface: Server side rendering is crucial for its performance and SEO benefits. Made choices around good performance.
- JavaScript framework: Migrating from Angular.js 1.6 to Angular 17 was a huge undertaking and there was no significant time savings by staying on Angular since Angular 17 is a completely different framework.
- Next.js: A few of our developers have experience with React and Next.js as a metaframework for building SSR (server-side rendering) applications was rapidly rising in popularity. We really desired the fast initial load and app-like navigation behavior that Next.js provided.
- Svelte: Reactive model was appealing and the programming model easier to understand as compared to React's, however the ecosystem is smaller than React and there aren't as many examples of successful e-commerce websites built in Svelte.
- Styling: The stylesheet was getting very bloated due to many classes being added over the years and being hard to remove.
- Tailwind CSS: Tailwind CSS was among the hottest CSS methodologies and its atomic CSS approach scales well for growing code bases.
- Styled Components: CSS-in-JS was something we were also looking at, but Styled Components was tied to React and runtime style injection was bad for performance.
- Performance-centric mindset: Read many performance case studies on web.dev and engineering blogs of other e-commerce companies, gathered a list of important performance techniques and processes:
- Set performance budget for each page (under 300kb)
- Run performance benchmarks as part of CI – before merging Pull Requests and continuously
- Lazy loading of non-critical components
- Lazy load below-the-fold contents
- Split JavaScript on page level instead of a single bundle (handled by Next.js)
- Use WebP format for images
- Host images on a CDN instead
- Adaptive loading of images so that mobile devices load a smaller image
- Consolidated duplicate JavaScript libraries (data-fns and moment.js), switched to
lodash-es
, removed all usages of jQuery - Looked at data to identify lesser-used features and removed them from the code, reducing JS size on the product details pages by 200+kb
- Search Engine Optimization (SEO)
- Used SEO tools like Ahrefs to continuously monitor SEO
- Worked with the marketing team to ensure marketing copy included important keywords as shown by Ahrefs
- Adjusted page URLs to include SEO keywords
- User experience improvements
- Single page checkout experience as opposed to two-page checkout to reduce clicking
- Reduced height of many UI elements
- Fixed checkout button that wouldn't be missed
- Payment improvements
- Analyzed Stripe checkout data and implemented country-specific address fields
- Initially only had one available payment method – credit card. Enlisted the help of data scientists to evaluate the popularity of new payment methods and whether they were worth adding. We later added PayPal, Google Pay, and Apple Pay payment methods as well
- User interface: Server side rendering is crucial for its performance and SEO benefits. Made choices around good performance.
- Solution evaluation
- Metaframework: Chose Next.js because it is backed by a company, there were many available resources, and has the largest community of all. React is also the most popular JavaScript library out there and also easiest to hire for
- Styling: Tailwind because it's a reliable and futureproof option (compatible with the direction React was heading)
- Monitoring and adjustment
- Rolled out the new website behind an A/B test while monitoring the performance and conversion rates over a period of 2 months.
- Countries which previously saw lower conversion rates experienced nearly 50% improvements in conversion rates.
Results
- Lighthouse score improved to 92
- Loading speed improved to 1.5 seconds
- Conversion improved to 2.5%
- Developer velocity has improved in the recent survey and it is now easier to hire people into the team because more people know React over other frameworks
Obviously the entire story is too detailed as a response for a single question; you definitely shouldn't be listing all the specific performance improvements you made, like a laundry list! Aim to achieve both breadth and depth by briefly covering each phase within the "Action" section and going deep into the parts that showcase the signals required for that specific scenario.
Possible nature of follow-up questions
As mentioned in our behavioral interview preparation overview, interviewers are encouraged to rely more on follow-up questions to really understand the candidate's thought process and motivations, which typically fall under these categories:
- Why do you think you did (insert action)?
- Why did you not do (insert action)?
- How would you do things differently in hindsight?
For questions on problem solving or drive for results, interviewers will most likely probe for questions to help them understand a bit more about:
- The source of the task/problem/objective (Understanding level of proactivity and initiative):
- Was the project or task initiated by you? And to what extent?
- Was the underlying idea from you, or just the plan to execute?
- How did you get buy-in from stakeholders to even work on it?
- Role and actual contribution by candidate:
- Was there a team involved with solving the problem or achieving the objective?
- Which actions were initiated on your own accord or contributed solely by you, and what did others do?
- How did you help the other members involved to achieve the goal?
- Prioritizing the problem or objective to achieve:
- Was there a reason why this particular problem or objective was prioritized to be addressed or done against others?
- Was there another problem or root objective that should have been more important?
- Was this problem not already resolved by another team in the company? Why did the solution have to be reinvented?
- Selection of appropriate metrics/objectives, and if they were measured post-launch:
- Was there any quantitative or qualitative target set, and how were they decided?
- How were they measured post-launch, and what was the result?
- If an issue was resolved, what measures were put in place to ensure the same problem didn't happen again?
- Using sufficient information to drive decisions:
- What types of research or data did you rely on to make decisions / choose a problem to resolve?
- How much time did you spend on research and gathering information?
- How did you balance between gathering information / planning and actually executing on the plan?
- Choosing the right solution:
- What other solutions were considered, and what were the pros and cons of each?
- Why was the final solution chosen?
- Who came up with the original idea of the solutions? How were they derived?
Sample problem solving questions and answers
The sample story above can also be tailored to answer specific questions by extracting relevant portions:
Can you tell me about a time when you had to use data to drive engineering decisions?
Elaborate on the section about conversion performance.
Can you provide an example of a time when you had to troubleshoot and fix a complex issue in a project?
Elaborate on the section about investigating the reason for poor conversion.
Can you describe a time you solved an engineering problem or achieved meaningful metric improvement?
Elaborate on the section about moving to server-side rendering and applying modern performance techniques.
Can you describe a time when you had to make an important engineering decision and how you decided between tradeoffs?
Elaborate on the section about choosing which frameworks to migrate to and the reasons behind the decisions.