Warby Parker Project Summary & Rubric Score

Rubric Score

Criteria 1: Report: Are conclusions clear and supported by evidence in all eight answers?

Score Level: 4/4 (Exceeds Expectations)
Comment(s): Your report and supplemental video presentation were outstanding. Your report presents your conclusions clearly and backs them up with evidence, using charts and visualizations effectively to help readers/viewers understand the data. Also, your video moves through the presentation at an appropriate pace, slow enough for viewers to process the material but not so slow that it drags. Great job! My main recommendation for further improvement would be to more consistently include context for the size/significance of the percentages you discuss. For example, in the Conversion Funnel slide (b), which discusses possible reasons why those 250 people did not progress from the quiz to the at-home try-on, you mention that 12% of those 250 people listed 'no preference' on the style question. However, there is no mention on this slide of what percentage of the 750 people who did move on responded with 'no preference' for that same question. As such, it is hard to tell how meaningful that 12% number is. Similarly, on the next slide, you mention 43% of those who did the at-home try-on but declined to purchase selected women's styles, but there is no mention of how many selected men's styles for comparison, making that 43% number seem less meaningful (while viewers/readers could compute the percentage selecting men's styles by subtracting the women's and 'no preference' percentages from 100%, it is generally best to avoid counting on your audience to perform mental math). Having these sorts of extra numbers for context/comparison can make the numbers you are seeking to highlight more meaningful.
Criteria 2: Query Accuracy: Do queries do what they were supposed to do?
Score Level: 4/4 (Exceeds Expectations)
Comment(s): Your queries consistently retrieve the appropriate information from the database. Well done!
Criteria 3: Query Formatting
Score Level: 4/4 (Exceeds Expectations)
Comment(s): Your code consistently adheres to the formatting guidelines, and you use comments appropriately to clarify the purpose of each query. Nice work! However, in your 'WHERE' clauses with more than one condition, I would recommend putting the 'AND' as the start of the line containing the second condition and indenting these subsequent lines one extra level. This formatting would be a little more consistent with how we handle JOINs and should make the code a little more readable. For example, lines 157-158 of your presentation could be formatted like this:
```
WHERE h.user_id IS NOT NULL 
AND p.user_id IS NULL)
```
Criteria 4: Understanding underlying concepts and terminology
Score Level: 4/4 (Exceeds Expectations)
Comment(s): Your written and video presentation demonstrate a clear understanding of funnels and their significance. Great job!

Overall Score: 16/16 (Exceeds Expectations)

This project was excellent. You demonstrated a solid grasp of SQL and of funnel analysis. Aside from my recommendations in the comments for Criteria 1 and 3 above, my main suggestion for taking your work to an even higher level is to note the limitations of the datasets and think about what further data or analysis would be useful. For example, in your presentation, you use both the number of pairs sold and the gross sales revenue as means of comparison for the success of the different pairs. There is a strong chance, though, that what Warby Parker is most interested in maximizing is profit, which we cannot currently calculate because we don't have information on the profit margin for each pair sold. This is information that Warby Parker likely has, though, and it could be worth noting in the presentation that this info should be acquired and used for further calculations. As another example, you observe in your presentation both that the 5-pair try-on results in a significantly higher sales percentage than the 3-pair try-on and that the quiz responses for color preferences do not match up perfectly with actual purchasing decisions. This brings up the question, are the 5-pair try-ons so much more effective because customers' preferences vary widely and they need the extra options, or is there in fact a strong but not necessarily intuitive correspondence between quiz responses and purchases that is causing the 4th and 5th pairs added for the 5-pair trials to sell better than those first 3? Could their system for choosing which pairs to recommend for at-home trials be changed in a way that could increase sales for the 3-pair trial? An examination of the glasses sent home for trials and their correlation with purchases could potentially shed light into these questions, but this is again data that we don't have, but Warby Parker likely either has or could get. Thinking about and recommending these types of opportunities for further study can help make your work even more valuable. But, again, your project is already in great shape. Well done!

jonathanherman1 / warby-parker-capstone-project