W4/W5 Product Teardown: Chefling

Note: Per Medium policy, I am reaffirming that I am not sponsored by nor advertising this product. Like any product that I am using in this Product Teardown series, they are only here for the purposes of education and discussion. A URL to the product will not be included.

It is crazy to think I’m already 5 weeks into my Product Management journey! I originally had a Product Teardown planned for last week. However, those plans changed when I (1) started to rethink my evaluation template and (2) found a product a little earlier in the development cycle. I’ll briefly summarize my thinking before getting in the product selected for this week, Chefling.

Why rethink the evaluation template to start with?

I had some time to reflect on my W3 teardown and the challenges I faced while evaluating RapChat. The first thing that stood out to me was the Market Evaluation segment towards the end of the teardown. It didn’t seem to fit quite as well, especially as I felt that I was rehashing some points and assumptions from the User Persona and Pain Analysis. RapChat was also a unique product in that it occupies its own section of the ecosystem, with very few 1:1 competitors.

On the flip side, without adequate market analysis and understanding of the service, the User Persona section at the beginning didn’t feel informed enough. So I restructured the template accordingly, with the intent to streamline the analysis and also delve a lot more into the User Experience. As the UX portion of the teardown continues to get longer, I wanted to add a small segment related to adding features towards the end — and try to think about how I would prioritize the backlog.

Let’s give this a shot.

This week’s product is Chefling, which has an enticing value proposition of being “your new AI kitchen assistant.” CrunchBase showed that Chefling has been active since 2016, with a lot of media focus in early 2018 where it started to raise capital and take traction. It has what I would consider a three-pronged attack: (1) recipe management, (2) meal planning, and (3) inventory management, which can be controlled by voice command.

After playing around with the product a bit, there were some challenges with the voice controls, so I decided to just focus on a non-voice navigated experience. Also, unlike the previous two Product Teardowns, I did have some familiarity with products in this ecosystem, as I myself want to have a little more control over how I eat, while having the occasional refrigerator straggler that needs to be used in a recipe ASAP.

Does this product solve a problem?

For sure. Products such as Chefling and its competitors solve one part of a larger problem. The two primary angles that I could identify comes from two groups of potential consumers, which could be inclusive. The first and what I imagined would be the larger of the two user groups are those who are participating in Dieting & Weight Loss. The second and smaller user group are people who are interested in kitchen inventory management and/or organization. From a user group dynamic, it is expected that there could be overlap between the two.

Tackling the personal challenges of Dieting & Weight Loss: although the problem itself is rather large (no pun intended), requiring commitment on the fronts of planning, tracking, and most importantly executing, at least this class of product would help facilitate a solution in the form of setup, which would drive more consistent engagement in the dieting plan. I took a closer look at some 2018 statistics (some are paywall gated) to paint a better picture of the Dieting & Weight Loss scene within the United States, seeing some pretty eye-opening numbers:

· The United States has a 31.3% Obesity Rate amongst adults

· 48% of adults believe that obesity is caused by diet

· Most popular reason for wanting weight loss is driven by overall health, appearance, and longevity

· Weight Watchers (a program where food intake is tracked/calculated) is the most popular diet trend in the United States

My favorite statistic came in the form of some pre-validation, from a 2017 survey. Specifically, with a noticeably disproportionate ratio of U.S. adults who at least expressed a willingness to use an app or product to track diet and nutrition. Especially within the two age groups that were 41 years or older (41–60, 60+ years), the refusal to use such a product was only 22% and 29% of those surveyed, respectively, with the majority of surveyed users at least showing a willingness.

Product Expectations

Hey! New section. The statistics in the previous section are a great way to transition into a market snapshot. Since the market of Weight Loss & Weight Management is lumped into one, I had to break it down to meal replacements, supplements, and services. I found a figure that pegged the online services portion of this market to be around $1.2 billion USD through 2018. Combine that with an increased desire to cook meals at home, and this sets up a strong snapshot and argument for a product like Chefling.

Based on this setup, I had several expectations of what I would see as features when I head into the User Experience portion of the Product Teardown. With a value proposition like that of Chefling, looking at the three-pronged offerings, I hypothesized that I would see the following:

· The ability to track on-hand grocery/pantry inventory

· The ability to plan ahead with some type of calendar

· The “ready to cook” recipes would be available with the ability to add personal ones

· Recipe discovery

· Cooking instructions

· Nutritional information tracking

The first four on that list were take directly from the website with the remaining two added as additional (and reasonable) expectations. If this product could help solve the problem of facilitating weight management by driving engagement, there would be opportunity in offering preparation instructions along with nutritional information. The nutritional information was key, especially with figures ranging in the ~30% region of Americans being on a specialized diet, the exact type that is driven by a need for nutritional information.

Pain Analysis & Market/Competitor Snapshot

Looking at the two hypothesized target user groups, how big is their pain? I would say that the magnitude of the pain could be classified as medium to high, depending on the user class. Given the statistics to back up the potential size of the Weight Loss & Management market, this is where the magnitude of the pain could be quite large, especially as surveyed users expressed health and life expectancy as some of the primary reasons for dieting.

I would classify the smaller user group, who just wants to be organized, as having a medium magnitude of pain. Although being organized is an outcome that these users want to reach, it is hard to justify it at the highest level when put side-by-side with health issues.

In terms of frequency, this would definitely be classified as a high frequency pain. Not only do most people/users eat multiple times a day, food itself is a perishable, which requires constant replenishment at various intervals to maintain needs. It’s why delivery services like InstaCart or Amazon Pantry have gained quite the presence in this type of space.

This would put the pain associated with the hypothesized users in line with a frequency > magnitude style of model, where both of these factors are on the high side. In this case, as the market size and presence of competitors would suggest, there is a lot of opportunity here to solve a real problem.

So where does Chefling fit into the ecosystem? I was not surprised to find quite a list of competitor products, where I was able to select a couple to do a quick Feature Table. Similar to RapChat, this combination of features landed Chefling in its own class of product, so I had to compare products on a tangential basis. I identified Tasty and Yummly as two very popular products that had recipe discovery and light (only a shopping list) inventory management ability. I myself am a user of MyFitnessPal, so that rounded out the four products. Taking a look:

The three competing products that were part of the Feature Table provide a good look into the strengths and weaknesses of what Chefling had to offer relative to the ecosystem and its value proposition.

Note: That is not to say that these aren’t features that could be added eventually; it is the current snapshot of the product as of today.

Most notably, there were a couple of categories that did not agree with my initial hypothesis as I had an extended navigation of the product. The most notable one was the lack of nutritional information within the product, which for me was a big differentiator for planning meals. I would love to see this as a future feature of the product, and considering this from a Customer Interview perspective, I wondered how many potential users would identify this as a desired feature. I tried to dig for a solid statistic that could represent the percentage of users that, on a regular basis, (1) prepared/planned their own meals, and (2) were on a macronutrient restricted diet or some type. Eating regiments such as Keto, Paleo, Whole30, and many more, all require some specific ratio of intake.

The other feature I was surprised to not see was the ability to curate personal recipes. In fact, neither Tasty not Yummly had this feature either — these products were all driven by existing recipes through various databases, with Chefling’s being draw from what I could tell was a series of blogs and other websites. Understandable for there to not always be consistent nutritional information, as an aside.

On the flipside, MyFitnessPal covered an entirely different array of features, overlapping with Chefling on a tangential basis. This product was more geared towards tracking intake overall, so there was not a focus on pre-existing recipes, cooking directions, or a shopping list. Interestingly enough, MyFitnessPal could be used as a meal planning calendar of sorts, and was nutritionally driven.

Other products existing in this ecosystem consisted of apps that were specifically restricted to a certain means (“Insta-Pot only,” salad builder, Holiday Recipes) or a certain feature (only a grocery list).

Time to get into the product!

The User Experience

I visited the Chefling website initially, which was done very well with an inviting image of a home-cooked meal. The initial landing page was very smooth and to the point, dropping its one-liner value proposition and offering quick download options for both Android and iOS.

Like previous products that I’ve taken a look at, I love getting straight to the point, and the website continued the trend by highlighting the three main features and the fact that these were tied together with voice control & synchronization. I headed to the Google App Store to download and install.

I was met with some simple, reinforced onboarding:

For a first time user, this is very helpful to reiterate what the features of the product are. I also discovered a feature I didn’t originally anticipate — the tracking of when foods were about to expire. I would find later on that this feature wasn’t quite as polished as desired, although it offers a lot of potential for in-depth inventory management.

The sign-up process was very easy, allowing for various options. I once again chose my teardown dedicated email address and signed up as Chef Mark.

As some additional information gathering, the product asked me for diet-specific restrictions and ingredient-specific restrictions. I naturally went with a type of diet that I would consume, as I tend to stray away from dairy and wheat heavy types of meals. When I got to the ingredient-specific option page, this is where I had a little bit of confusion:

I would find out later that this would be a screen for eliminating ingredients, not selecting them for inclusion. This is interesting to me from a UI perspective, where I would have liked the “Dislike Recipe” line to be a tad more apparent, or at the least not conflict so much with the “…recipes you’d like” line. Especially because the first screen was a positive reinforcement for a lack of an element, whereas the second screen was a negative reinforcement for a lack of an element.

I was met with a “What’s New” splash upon entering the app, where I was met with the main screen, consisting of quite a few navigation options. After looking around, I found my profile at the top of the UI, where I was able to correct the mistake I made from the ingredient restriction questions.

I was ready to delve a little into the offerings of the product, heading back to the main page and looking at the three tabs on the Recipes Page: Favorites, Read to Cook, and Discover:

As expected, the Ready to Cook feature would require me to manually add ingredients to the digital “pantry” for use, so I decided to tackle that first while looking forward a bit to what I was really excited about: the “Discover” feature. I input a few simple ingredients in the digital pantry, which consisted of manually entering some chicken, onions, green beans, and apples.

I had not used recipe apps like Chefling prior to this Product Teardown, and I was happy to see the app break down what my available options were by category, while offering some visual aid. At the top of the page, I did get to take a peek at Chefling’s monetization model, which would be fueled by the inclusion of some ads.

However, unlike some other products that I’ve used where the ad is completely irrelevant to what the app is trying to accomplish, I found this to be a great way to include a monetization model based on a value proposition. Have these X ingredients and want to make Y recipe? Here is an ad where you can get it, and oh by the way, it goes straight to the shopping list that’s part of the product.

I did get a little skeptical regarding some of the category selection, which put Onion as its own category even if it was not the featured ingredient. Specifically, this presents an opportunity down the road for fine tuning the search capabilities. It’s completely understandable that if I have an onion, that it could be part of this hamburger dish pictured above — however, how much does an onion comprise of a hamburger dish? Maybe 5% of it? So what would the search algorithm do if I had, say, pickles? A possible outcome that Users could desire would be relevant search results. Having to go buy beef, which is a featured ingredient that most certainly would take up 50%+ of this dish by volume, as a result of this search, conflicts a bit with the value proposition of “Ready to Cook.”

I also wondered how the categories were prioritized. For example, the Breakfast category appeared where the only dishes available were green bean dishes. Then the Egg-free category appeared shortly after, which was not a diet-restricted selection that I had made prior. All things that could be tackled with small product teams in a development scenario.

For now, onto the Discover option!

At an initial glance, I put in a requirement of chicken-based recipes, which required baking, and could be completed in under an hour. Similarly to the “Ready to Cook” option, I was met with quite a few search options — these were not formatted like the previous section and instead presented as a large pictured list, which would align with the concept of idea discovery. See the myriad of options, select any option.

The first recipe looked pretty good! I selected it and immediately I could see that I only had 3 of the 11 required ingredients (bummer), along with some options on how I wanted to manage the recipe, including the option to favorite the recipe (and add it to the home screen).

The other options allowed me to add the missing ingredients directly to my shopping list or even put the meal into my plan. Oh, and cooking directions. Where did those come from anyways?

As it turned out, the cooking directions for this specific recipe came from a blog on the web, which contained its own embedded ads. This I was a lot less of a fan — the page was littered with ads (two overlapping ones in this case) and were not really related to what the product was trying to accomplish. Comparing this directly to competitors like Tasty and Yummly, which kept their directions within the app, I had concerns about safety and navigation. It seemed to me that Chefling was scraping/pulling recipes from various sources online, so the company itself is not able to control what users are exposed to if something about an external site is changed or updated.

I did like how the shopping list automatically checked items that were not accounted for. That said, there could be some difficulty working things out like generic “Herbs” versus substitutes or equivalents, which again like the previous section would have an effect on search engine optimization. I took a closer look at how the shopping list feature itself was integrated, starting by adding the missing ingredients from the dish I found.

Another ad, although this one was not really related to what I was trying to achieve with the shopping list (so how are they picked out?).

Oh hey, what’s this bag icon do?

I typed in my zip code and found that the app was integrated with nearby stores so I could get quick access to missing ingredients. Although it wasn’t quite the InstaCart that was integrated into the Yummly app, in theory this satisfied the same requirement. While there were definitely some opportunities for a better set of results from the search engine, especially with the way products were categorized or named, scrolling through the options at my local Walmart I could see that the app got it close. I imagine that at this point, there are still some difficulties with optimizing search results.

How about other ways to use the shopping list feature? After all, there were several other options aside from exporting it out of an established recipe.

How about from a receipt? I took the latest one I had and after struggling a little bit to get the right lighting and angle, the app took a pretty solid picture and started extracting data:

This was incredibly close! The only thing that the app understandably didn’t pick up on was the EFBG/PEM item — something I honestly do not remember (maybe some almond milk?) and might have been a store specific item. Otherwise, most of the ingredients transferred over very well. I gave a shot with the barcode scanner as well, where the product easily identified a can of black beans. The result was a fully populated shopping list.

Now, what about the pantry side of things?

The pantry feature is setup similarly to the shopping list, with the ability to add or remove items as necessary. I was greeted with an expiration alert for my chicken, with a default time set upon entry.

While it was good that Chefling suggested potential recipes that I could be using to consume the chicken (or other expiring items in the future), I did notice that there was an opportunity for improvement in how the information for each item in the pantry was stored. There wasn’t a unit of measure available for this particular item, which was how a lot of ingredients were set up — definitely something to add in the fine tuning stages.

As one final part of the UX exploration (at this point I’d be working around the app for quite some time), I decided to throw everything I had on hand and see how what was available for me to cook.

Although it was easy to add ingredients in this stage, I had a little bit of struggle here from an ingredient identification standpoint. There were some items which were repeated, like adding in oatmeal, while other ingredients I used weren’t integrated in the app yet, like kimchi. Initially, while trying to be thorough with what I was adding to the pantry, including basics like salt and pepper, I headed back to the “Ready to Cook” screen, where I was met with an array of chicken dishes and 13 recipes which did not contain chicken at all.

This went against the search results earlier, where some beef recipes were included despite not having any on hand. It seemed like the addition of detail with as many condiments and seasonings as I could had negatively affected the results. I also found out that just adding plain water could pull another five recipes into the results that did not contain chicken.

Before finishing up the UX navigation, I quickly flipped through the available recipes to try and understand how Chefling pulled results.

It’ll remain a mystery for me. For example, the above recipe was generated only because I added salt, a very low percentage component of the recipe. I had peanuts on-hand, however they were not classified as roasted unsalted peanuts so it didn’t detect that as an available ingredient in the app itself. The ingredients were actually pulled from the Tasty website, with what appears to be… a small error in translation. It was interesting to see that some lines of ingredients were pulled directly from the Tasty recipe, like the honey and canola oil, while the app had an added unit of measurement for the peanuts.

This was definitely one of the longer UX walkthroughs that I have done. Although I didn’t get to sample the voice portion of the product, I still got to see the many features of Chefling through traditional navigation.

Generating a User Persona

A familiar section from previous teardowns, I moved this question to follow the User Experience, with the intent that fully moving through the app would help confirm or deny my original assumptions and expectations about Chefling. So, what could be one User Persona associated with Chefling?

I went back through the statistics in the Market Analysis portion, as well as how I felt about the UX, to generate one potential User Persona. Looking at the Pain Analysis, I tended to favor what I perceived as a higher pain, the Weight Loss & Management side, rather than the organization side. The one piece of statistical information that stood out to me was part of the 2017 survey regarding potential engagement of using an app to track diet and nutrition — there was a significant increase once the user group reached the ages of 41–60. This allowed me to construct: Conscious Charles.

I wanted to move through the alphabet a little more when generating these User Personas. However, the one word that resonated with an age group that was just getting into the diet and nutrition tracking and realizing the importance of what food is taken in, was Conscious.

Conscious Charles fits a potential User Persona of Chefling. He’s finally at an age where he can’t just eat whatever he wants and some recent doctor visits reveal that he needs to start eating better. Charles is realizing that he needs some serious help getting everything into shape. He’s used to takeout food, his refrigerator is a mess, and with the fact that it’s still very early in the year, he sets out to make this a resolution.

Chefling’s features are just simple enough for Charles to look up potential recipes, and refill his pantry or shopping list with a scan of a single receipt. The product alerts him when some of his meat and produce are about to expire as well. Since he’s a pretty busy guy and works late, occasionally he’ll use the “Discover” feature to find something that he can prepare quickly.

This plays very well into the Empathy mapping as well — using this User Persona, the biggest focus should be from a “say and do” perspective. At this point in life, what matters the most to Conscious Charles? He’s part of a demographic group that should be looking to improve their health and longevity for years to come. Chefling is a product that can help him do that.

If I were the Product Manager…

This is a new (and final) section where I try to take a look at the product if I were the Product Manager. Based on the information that I’ve uncovered through the Product Teardown thus far, I was able to generate a SWOT Analysis for Chefling:

With a lot of attention towards the Weaknesses and Opportunities, it is clear that Chefling has a strong basis of features and an excellent value proposition. However, to truly be a strong differentiator against its competitors, direct or tangential, there are some areas of the product that needs to be prioritized, many of which have been identified during the UX navigation.

Most importantly, in the Threats area I specifically list that overlapping features from multiple tangential (or direct) competitors could affect user retention. What I mean by this is a scenario where Chefling’s development has stagnated or isn’t strong enough to differentiate itself from a combination usage of something like Yummly and MyFitnessPal.

For example, even if Yummly doesn’t have a planning calendar, MyFitnessPal effectively has one, which means that these two products together effectively covers the features of Chefling. MyFitnessPal even offers additional features. Given that these apps are already part of the established market for this ecosystem, what does Chefling have to do to break through and attract users? Is the voice command feature going to be enough to draw a user base away?

So what are the backlog features that I would focus on as a Product Manager?

I may initially shelve the idea of getting the voice command feature online, at least until the user flow has been optimized. The following three items, with justification, are what I would move up on the backlog for Chefling:

1. Recipe Search Consistency. This issue came up a couple of times during the UX navigation, from entering available ingredients to discovering new recipes. It was not clear how the recipe search process prioritized ingredients, sometimes returning a dish like the peanut butter shown at the end of the UX navigation — only salt available in a recipe with three other major ingredients. Should this even return a result if this is the case? Additionally, while it is understandable that recipe searches are very complicated, the feature could immediately benefit (small win) from a decision made around whether basic items like water, salt, and pepper, should even be affecting search results to begin with.
2. Ingredient Streamlining. During the shopping list exploration, I encountered ingredients like oatmeal which were entered into the database twice. Some ingredients did not have units of measure, while some ingredients didn’t exist at all. Other ingredients, like unsalted peanuts/regular peanuts/roasted peanuts existed in quite a few forms, which invalidated some potential recipes that could have been used if the search algorithm was more flexible with ingredients.
3. Nutritional Information. I struggled to add this one to the top 3 initially. However, after some thought, nutritional information is something that is fairly important in terms of differentiation — in this case, the ability to stay on par with competitors. While the Tasty app doesn’t have nutritional information, looking at Yummly and MyFitnessPal as potential competitors, these apps already include such information. The one positive in this case, is that based on prioritization, as the ingredients continue to be streamlined, the ability to add nutritional information should get easy over time.

With these three items on the backlog being constantly worked on, it sets up Chefling quite well with some distant preparations for a Personal Recipe Creation feature. It wouldn’t make sense to focus on this feature initially, especially as the ingredients and nutritional information associated are somewhat incomplete in the current state.

That said, I consider those three items listed to be some fundamental building blocks to drive Chefling’s ability and growth within this ecosystem. Streamlining the basics like search and ingredient information will remove a headaches while refining existing features or implementing new features.

Looking ahead, one long-term goal would then be the consolidation of recipes and directions onto a Chefling-held database. This also was mentioned in the Threats section of the SWOT, especially with regards to relying on external websites and factors to be the core content of the product. As the product directed me to an external website during the UX navigation portion, it led to quite a bit more ads and some questions on the security of using the product itself. While it’s understandable that this may be a lot of work, which is why I identified it as a long-term goal, it goes very well hand-in-hand with the streamlining of the fundamentals when it comes to Chefling’s content.

Final Thoughts on Chefling

This Product Teardown on Chefling was a great opportunity to explore a product that has a great basis to offer along with a solid value proposition. Although many of the features still have opportunities for improvement, a move through the UX navigation and market analysis puts Chefling in a position where it can definitely succeed. It exists in an ecosystem where there are already plenty of direct and tangential competitors, and it does have to address some big challenges if it is to differentiate itself.

The solution that Chefling offers solves pains held by what I originally hypothesized were two major user groups, those in Weight Loss & Management and those who were interested in kitchen organization. After experiencing the product, I did not shift much from that perspective and solidified the creation of the “Conscious Charles” User Persona. Especially with regards to the Weight Loss & Management user group, it is a high magnitude, high frequency pain in a growing industry.

Performing the SWOT Analysis for Chefling, I identified the need for streamlining recipe searching and ingredient information. This is not only because it would help in the short-term with the basic features of the product. This is because it sets up building blocks for so much more down the road within this ecosystem. Although the voice command feature had some functionality, I personally had some challenges and did not prioritize it as much when looking at potential backlog features to elevate.

Overall, I’m excited to see how Chefling develops. It has the potential to be a very strong product, thinking far ahead to what it would be like in the late development stages with full capability and integration.

This concludes my Product Teardown of Chefling.

Let me know if the new evaluation template makes sense! Thank you for taking the time to read.

-Chef Mark