

Product management is tough to describe. I summarize the role as FSO – Figure Stuff Out.
It can include concepting, research, pricing, strategy, business development, marketing, UX design. In my roles at Amazon, Microsoft and Aol, despite having “product” in my title, the focus of the roles was very different. In large companies, it can even vary from team to team. Nascent products often have each person span more areas, whereas established products have more discrete roles.
At startups, one person usually covers most or all of these roles. The specific division of labor depends on the skills of the team and the needs of the company. These shift rapidly as the company grows or pivots.
My product philosophy is different from the prevailing view in Silicon Valley. Products should be driven by how people live and work (or will work), not purely based on technology. There are many products that have failed because they were driven by a virtually exclusive focus on technology. Google Glass and Google Wave are notable examples.
Of course, we need the technology to enable these ideas.
Technology + Psychology = Great products.
I ran a series of classes on product management. This 11-week course covered every aspect of product management. Even if you are focused on a specific area, like research, it helps to understand how pieces work together.
These decks provide an outline of what was discussed in each session. The live sessions contain more deal. Check back here to see when the next series will run.
I work with companies on a consulting basis on all of these topics. (Both startups and large companies.)
I also run 2-day, customized corporate training sessions. If you’re interested, please email me at inquires@agrawals.org.
Are you building the right product? It’s an important question whether you are a startup or a big company. Good research can help guide you. Doing it incorrectly and you’ll go down the wrong path.
There are two basic types of research: qualitative and quantitative. Qualitative generally involves asking people what they want or their experiences with existing products. Quantitative using hard numbers from your users.
Quantitative research can help you answer questions like “What features do I need to add to my product?” “What features can I remove from my product?” “How is my user base generating revenue?” “Where is there fraud and abuse?” (There is some overlap; I’ll do a separate post on qualitative research.)
Some caveats to look out for when doing quantitative analysis:
These are some of my favorite things in quantitative analysis. This is by no means a complete list.
A/B testing
This is commonly used to test different messages or designs. Two variants (A and B) are presented to different users. Marketing emails commonly use A/B testing. Take a small portion of your subscriber list and send one subject line to half and another subject line to another half. With the data on open rates that you get from these emails, you can send the one with the better conversion to the rest of your list. There can be more than two; you can have A/B/C.
1% testing
This is a variant of A/B testing. It’s commonly used to test different features, especially in complicated products or products so well established that you don’t want to change the experience overnight.
Take Facebook’s News Feed. This is a product that is used by billions of users around the world. Adding a new feature without testing can cause a lot of grief and negative feedback. Before you roll it out widely, you present the new feature to a tiny percent of the user and track how it performs. Do people use it? How often do people use it? Does it add or subtract from other features people use. (I call it 1% testing, but in Facebook’s case, it might be 0.001% testing.)
Market segmentation
One of the challenges with data is that averages can mask important differences. You can dig into data to identify segments that you want to go after. If you’re running a credit card business and find that 15% of your overall spend is travel, that tells you one thing. But when you look deeper, you find that a group of customers spend $50,000 a year on travel. This might lead you to create products for that lucrative customer.
You can also use data to figure out who your profitable and unprofitable customers are. In many products, you’ll find that some customers are unprofitable. They could be doing too many returns. (E-commerce.)
Fraud/abuse analysis
Detecting fraud (illegal behavior) and abuse (legal behavior but not within your business model) is a great way to use data.
I worked for a company that sold long distance calls. When we looked at the usage data, we found that we had a very large amount of usage to Tuvalu. Given that it’s a tiny nation, this didn’t make sense. A closer look found that there was an error in our rate tables and we were selling something for 10 cents that cost us $2.00. As you’d expect, people from Tuvalu told each other about it and we became the calling service of choice for them. (Some of the details here have been changed.)
Another use case is finding the outliers in all-you-can-eat plans. Think about cell phone data plans. In the AYCE model, some customers might use 1 GB of data and others use 100GB. Your business model and network capacity is based on average usage of 5 GB of data. The 100GB user hogs capacity and slows things down for everyone else. With data, you can develop new policies: the * that says data rates will be slowed down after 25 GB of use.
Search analysis
Looking at what people search for but you haven’t delivered is important to product planning and improving the experience. After all, they came to you for it.
Let’s say you run a ride app. When someone launches the app, they might be an area where you don’t offer service. Tracking those requests gives you insight on markets that you might want to look at when developing expansion plans.
It’s also a way to improve the product to suggest alternatives that the user might want. If someone is in New York City and searches for “In-N-Out,” you might respond “There are no In-N-Out burgers in New York City, but here are some McDonald’s.” Just kidding. I’d probably return Shake Shack, but In-N-Out is so much better.
Cohort analysis
The key to a successful business is that lifetime value is greater than customer acquisition cost. (Often written as LTV > CAC). You want to make sure that, on average, you make more from customers over their lifetime than it costs to get them.
Look at the customers that signed up for your service 1 year ago and how much they spent and when. For customers that sign up today, can you use the historical data to model what the new customers are likely to do?
When looking at data, you also have to weigh the cost of the analysis against the value of the data. If you’re using data to analyze how people navigate through your site, it may be sufficient just to track data on a small subset of users. Adding too much tracking can add to latency in your site or app.
Also, if you’re trying to decide whether or not to implement something that will take 2 days, it doesn’t make sense to spend 2 weeks to build a system to get the data.
There are a variety of tools you can use for quantitative analysis, depending on what you’re trying to get at. Marketing tools like HubSpot handle A/B testing for email campaigns. Google and Facebook have their own tools for ad performance analytics. Google Analytics and Adobe Analytics allow you to analyze user behavior. For complex feature-level data, you will likely have to create your own database and run SQL queries against it.
COIVD-19 Caveat: For most businesses, I don’t recommend doing quantitative research based on data beginning in March 2020, unless you’re using it to compare the impact of COVID-19. If you try to project based on data from March 2020 on, you’re likely to over or underestimate behavior post COVID-19.
Grammar trivia: Data is plural, not singular. (Plural of datum.) It’s one of those weird English things that doesn’t seem right, but is. Like how a person who runs a restaurant is a restaurateur, not a restauranteur.
In week 2, we’ll focus on pricing. Class is at 5 p.m. PT on June 3, 2020.
Cost + margin = Price has been pricing for a long time. But as products have gotten more complex, pricing that way often means leaving money on the table — either by pricing too high or too low.
Often, you have a pricing framework and you let algorithms do the rest. The biggest, more profitable companies — Google and Facebook — do it that way.
For this week’s class, Adam Crouch will join us. Adam is VP/General Manager for new markets at Poshmark. He previously worked at Volvo and Walgreen’s. He has a lot of experience in pricing with different models.
If you’ve already registered for previous classes, you should have a calendar invite. If you don’t see the calendar invite, please DM or email.
If you want to register, please fill out this form. As always, classes are free. Drop in anytime you look.
It’s 2020’s Holy Grail: the coronavirus vaccine. Large parts of the country and the world are shut down in a deadly pandemic. Scientists around the world are racing for a vaccine to end the suffering and re-start the economy.
Let’s say you found it. After 9 months of hard work, you have the answer to the world’s problems. What do you charge for it?
(Unlike most strategy and pricing questions I pose, there isn’t what I believe to be a “right” answer. This post is a framework for you to think about pricing in general. In all likelihood, you’re not developing a vaccine. But, if you do, I’ll help you think through pricing in exchange for a lifetime supply of your vaccine.)
Some factors to consider:
Consumer value
No doubt the value is here. At a country level, we should be willing to pay $1 trillion — we’re spending more than that in the current bailout and recovery packages, with an unknown amount of misery to go. From an overall health of the country and the economy, the government should write a big check that covers all 350 million people.
It’s when it comes down to consumer pricing, it becomes a lot tougher, In this scenario, I’d be willing to write a check for $100,000 to solve this for me and my spouse. Some would pay more; most could only afford a fraction of that. In economic terms, this is largely a price inelastic good. But I don’t get any meaningful benefit unless the bulk of the population is immune. (You can’t go to restaurants, go to a bar or get a haircut.) The price needs to be set so that the average person can buy it.
This is different from a drug like Sovaldi, where a 1-month, $28,000 treatment can cure you of Hepatitis C. There aren’t dependencies on the behavior of the rest of the population.
Cost of development
Cost + margin is a common (and lazy) way to develop prices.
In the case of drugs, the first pill costs you $2 billion and the next one costs you 5 cents.
The cost of R&D is less material if you’re Pfizer than if you’re a biotech startup. Pfizer might want to do it at a loss for other reasons; a startup that may only have one big drug in its lifetime needs to price differently.
Delivery timeline
One of the challenges in pricing is that people assume that if something takes longer, it is worth more. Clearly it involved more effort. So it’s “fair” that you charge me more. That’s how most industries work and how employees generally work.
In this case, a solution today is much better for society (and the individual) than a solution 6 months from now.
I would pay a higher “delivery” fee for my meal if it showed up immediately in my apartment versus waiting 45 minutes for it. But that’s not how most people think about these things.
Frequency of use
If one use cures or prevents the disease, then the pricing should reflect the lifetime value because you have to get it all up front. See the Sovaldi example earlier.
On the other hand, if you need to use it weekly or monthly, you can charge less because you have a long term revenue stream.
Brand impact
Depending on your pricing strategy, you could have a positive or negative impact on the brand. Some people say that if Pfizer discovered a vaccine, they should give it for free immediately to everyone because they would be remembered forever as the savior of the world.
But does the brand matter? In the case of drug companies, it really doesn’t, for two reasons. First, very few people know what companies make what drugs. Without Googling, what drugs does AbbVie make? Even if you do Google it, the first page of results don’t tell you what it makes. (You have to click on a link.)
The second is that pharmaceutical companies have a monopoly on the branded drugs they make and someone else gates what you purchase. Your doctor is going to prescribe whatever drug the hot pharma sales rep who took him to lunch told him about the drug that is the best for your condition.
There will be some initial PR blowback, but in the long term, it won’t matter. This is partly how doses of insulin cost hundreds of dollars – and the price keeps going up – despite the fact that the patent for insulin was once sold for $1. Eli Lilly, Novo Nordisk and Sanofi just don’t care what you think of their brands.
The brand impact matters more if you’re Target, Walmart or Pepsi.
Regulatory impact
For a coronavirus vaccine, this is probably the biggest constraint on pricing. Charge too much and the government may pay the bill to get the pandemic under control and then start probing every other aspect of your business.
In some cases, some governments will say, “Screw you and your patent. We’re going to make it ourselves.” This is especially true of developing countries. In a pandemic, this isn’t a hard decision to make.
Impact on other products
“This is an important life-saving drug and everyone should get it for free.”
Well, that may be true. But there are a lot of life-saving drugs. Do you give all of the life-saving drugs out for free and only make your profits on the quality-of-life drugs?
Have you undercut your entire business and the way people think about healthcare? (Leave aside the issue of whether pharmaceuticals should be a for-profit business.)
As I said at the beginning, this is not a guide to pricing your coronavirus vaccine. But these principles apply in pricing most things. As tech becomes a much deeper part of society, we’ll have to pay more attention to regulatory impact than we have so far.
For most products, competitive pricing will also matter.
There has been a lot written about how to be a great product person. The skills for being a product leader are different. Here are some of the things I’ve tried to do in my time as a product leader. (Many of these skills apply to leadership in general.)
These are, of course, general guidelines. You need to be aware of the specifics of your situation. If you’re in a highly regulated industry, it may have serious consequences if you fail; it’s the product leader’s job to understand that.
You’ve launched your product. You’re getting some adoption. People seem to like it. You start to get some feedback. TechCrunch does a write up on your site. You get lots more feedback.
Sounds great, right? In many ways, yes. But feedback can destroy your company, if you let it causes you to thrash in what you are doing.
Your earliest adopters tend to be the most geeky, the techies who probably don’t represent your target customers. (Unless your target customer is techies.) The feedback they give you isn’t going to matter to the general market.
I heard Stewart Butterfield, the founder of flickr, talk about this early on. Some of his users wanted that ability to geotag photos by uploading tracklogs from Garmin hiking GPSes and syncing their timestamps with the timestamps on their pictures. Great idea. I loved it and would’ve used it. But me and 10 other people. It turned out not to be necessary in the long term because cell phones (which are the source of the vast majority of pictures online) automatically embed GPS information.
In the short term, geeks like me found tools like GeoSetter and iTag to embed GPS date into EXIF fields on photos.
There are a few other problems with customer feedback:
In order to make the most of the feedback you get, you need to organize it. See how much of a certain type of feedback you get. Group similar feedback. Have your product and business teams decide where it falls within the business goals.
This is especially true of business-to-business companies. You’re having one-on-one interactions with the buyer. He’s giving you some interesting ideas. Sales tells you that if you had “one more feature,” they’d close the deal.
Wait. When there are a small number of big customers, you risk becoming a captive development arm. You definitely don’t want to be that. It makes it easy to lose track of your other customers. If your product becomes too specialized for a specific customer, it can hurt your valuation.
You want to take in feedback and harmonize it with feedback you get from other customers and potential customers.
It can be hard when you’re working closely with someone to say “no.” Your sales team might keep nagging you for that feature. (In a future post, I’ll discuss how to deputize your sales team and sales engineers into mini product managers.)
Stay strong. Make sure what you’re doing aligns with your roadmap and will appeal to other customers.
There are somethings early customers will ask for that are no-brainers. Examples are GDPR compliance and SSO. You don’t have to do these right away, but be sure to have an answer prepared on how you will solve these problems.
Yesterday, I sent some feedback to the CEO of a company I’m looking to invest in. It’s an incredible feature. I know it will help them sell to new customers. Of course he should take my feedback: I’m a potential investor, reasonably well-known writer, have decades of product experience and I run UX quizzes on Twitter that people seem to value.
His response?
Sounds like a good idea.
That’s something that we can think about adding to the system. Thanks for the feedback.
User research is an important part of the product design process. It can help you make sure you’re building the right thing for the right people and to continue to learn from what you’ve built.
I divide companies into two types: those that are driven by users and those driven by features. User-driven companies pay intense attention to the needs of their users and build products for them. Feature-driven companies focus on what they can do with technology and build those products. You don’t want to be the second.
Needs can be expressed or unexpressed. Some needs that users express may actually be counter to what they want. Important needs may go unexpressed because users don’t know exactly how to express what they want, or the research tools are poor.
There are many types of tools that you can use for research: usability testing, focus groups, ideation sessions, ethnographic research, user feedback, survey research, eye tracking, etc. Too often product managers confuse the roles of these tools and end up with bad interpretations.
Research can cost from nearly nothing to tens of thousands of dollars. In this series, I’ll go through research tools I think provide the biggest bang for the buck for startups.
Today’s topic: usability testing.
You need to assess whether people can understand what your product will do. Ideally, you’d do this shortly after the user flows have been built. Bring people in 1 by 1 and watch them accomplish a series of key tasks. Have the designer and product manager sit with the interview subject.
I like to just give the subject a task and see what happens. (After all, no one is going to be sitting with them while they are actually using the product.) “Order a large pepperoni-mushroom pizza and have it delivered to your house.” Then watch as he goes through the user flow. Have the interview subject articulate their thought process as they perform the task. Where did his eye go first? What was the overall flow?
Only when the subject gets stuck should the interviewer intervene. “Where did you expect to find it?” “Maybe try scrolling further down the page?”
Although ordering a pizza is a relatively simple task, there are a number of steps: finding the restaurant nearest your house, selecting your pizza, adding toppings, special instructions, delivery information, payment, etc.
Some people like to do the testing one step of the process at a time. I prefer to do it all in one pass. There are two primary reasons:
Once the first pass is done, you can repeat the test step-by-test and ask for detail on what was confusing or not.
Usability testing can be done with static images linked together, but I really prefer doing it with live or close to live code.
You can also test a couple of different flows.
It’s important that the subject is told at the outset that he is not the one being tested; it is the user interface that’s being tested. There is no right or wrong answer.
Whom you choose to participate is key. On HBO’s Silicon Valley, the initial response to the compression tool was amazement. All of their friends loved it. Except for Monica, who hated it. Then, we saw the user interface:
The people who loved it, the initial test group, were friends of engineers who were also engineers. They loved this type of interface because it gave them control over every little detail of the compression. Most users don’t want this.
You want to get people who don’t do tech for a living, if at all possible. You want to get people who are as close to your target audience as possible. If your product is designed to be used by senior citizens, don’t have millennials do the testing. (In this example, you’ll miss things like font sizes.)
Usability testing doesn’t need to be a formal process and doesn’t require special equipment.
You can recruit test subjects via craigslist. $75 to $150 for 40 minutes is a reasonable amount to pay. If you sell a product, some of this could be in service credit. For example, a $75 AmEx gift card and $50 to use on your app. I’m also comfortable asking friends and family, if they would be in the target user group.
Usability testing isn’t survey research; you don’t need to interview 1,000 people. With ten people, you should be able to identify the biggest obstacles in your interface.
The coming of self-driving cars will redefine the urban landscape, in the biggest change since the creation of the Dwight Eisenhower interstate highway system.
Interstates took motorists and money from roads like Route 66 and shuttled them through high-speed lanes with occasional exits. Motorists services sprung up around these interchanges.
Much of the impact from the interstates affected rural and suburban areas.
Self-driving cars will have their most profound effects on urban cores, where density provides the greatest value. (In rural areas, individual ownership of vehicles is likely to continue at a greater rate than in cities.)
These changes will not happen overnight or all at once. These will take decades to take effect and timing will vary by city.
Developing the best outcomes will require close cooperation between cities and the private sector.
Parking
Parking is the bane of an urban dweller’s existence. In dense urban cores like San Francisco and New York, reserved parking spaces can cost $450 a month or more. Parking meters have strict time limits, costs and fines for violations. “Free” street parking can require circling for 30 minutes or more.
Parking is needed because most cars sit idle for much of their existence; cars in use, by definition, aren’t parked.
With fleets of self-driving cars, users will be able to order cars on demand. Instead of owning a car that is idle more than 90% of the time, riders will be able to push a button and have cars come to them. As soon as a ride is complete, the car can be repurposed for the next rider.
This should lead to a decline in the demand for on-street parking, parking lots and parking garages. Demand for parking enforcement officers will also plummet.
This reduced demand will also affect city revenues. In San Francisco, parking is a $130 million business between parking fees and violations. Parking fines alone are projected to be $87 million in FY2017. (Of course, collection, maintenance and enforcement expenses will also drop.)
Space freed up from commercial and public parking lots can be reallocated to higher value activities like retail and residential, potentially increasing the city’s sales and income tax base.
This can also have positive environmental effects. Heat-absorbing asphalt can be replaced with green roofs.
Traffic
It’s estimated that more than a third of traffic in urban cores can be attributed to motorists searching for parking. Much of that traffic can be eliminated by substitution of on-demand self-driving cars.
Self-driving cars will also reduce traffic congestion in two other ways: not blocking the box and fewer accidents.
During rush hours in San Francisco, it can take 30-45 minutes to travel 1 mile on a street like Bush St., which funnels traffic to the Bay Bridge. Some of that congestion can be attributed to vehicles that block the box and induce gridlock. As self-driving cars take to the road, they can be programmed to avoid blocking intersections, smoothing the flow of traffic.
Accidents are a major source of traffic congestion as they reduce traffic flow capacity. The rubbernecking effect exacerbates these problems. Self-driving cars should dramatically reduce the number of accidents. When there are accidents, time spent on roadside investigations (which increase the rubbernecking effect) can be reduced based on access to data collected by vehicles.
Urban architecture
With less space designated for parking, space can be more efficiently utilized. Homes no longer need to allocate 1/3 of their space to fitting a car.
Instead of lots scattered around the city, parking can be relegated to staging areas. For example, commuter buses in San Francisco currently park during the middle of the day outside of the core. They come back into the city during the afternoon rush to take suburbanites home.
Infrastructure
Beyond simply autonomy, the sensors in self-driving vehicles can be used to collect and transmit data in real time that can help to improve infrastructure for all motorists.
Potholes can create traffic hazards and cause wear and tear on cars. Pothole repair often relies on motorists to report potholes. With self-driving cars, pothole locations can be detected and sent (along with pictures) to a city’s public works department. Instead of prioritizing repairs to potholes near the most vocal residents, they can be prioritized based on severity and degree of impact to the most motorists.
Signals are timed based on historical traffic patterns, not actual traffic conditions. Weather, special events and detours affect traffic patterns and can create suboptimal traffic flows based on signal timing. Data from sensors on connected cars can be used to optimize signals.
Safety
Sensors in connected cars can be used for other purposes that benefit society. For example, data from cameras can be used to identify suspects in Amber alerts or find dangerous suspects in emergency situations.
Challenges
A massive change of this sort doesn’t come without a lot of challenges, some technological, others societal.