When I was at Microsoft there was this big push back in 2002/2003 for us to be Customer Focused. It came from the top (Steve Ballmer), so it must be done. Back then, I thought I understood what that meant. I've read customer feedback, I participated on Usability Studies sessions, I pushed my developers to do the same, etc. So I must have been Customer Focused...
Then I left and about 9 months later I released the Alpha 1 of Sampa. As always I tried to be Customer Focused. But this time it was so much more, I was deeply worried for the sake of my business that I would understand, listen to and deliver on my customer needs. That was so much more than I could ever have realized at MSFT.
Enough stuffing, here is what we do in Sampa today:
#1 - Publish contact email everywhere
Sampa customer feedback email is feedback@sampa.com. I published many times on this blog, on the Sampa blog and it is easy to find on our website (www.sampa.com, click on Contact on the footer.) I do get a lot of spam because I put the email publicly, but the benefits of having an easy email is more important than the worry of receiving a couple hundreds of spams per day. BTW, we also have enabled a lot of aliases for our feedback email, because sometimes people just try random email addresses, so, all of these work as well: bug, bugs, sampa, support, contact, sales, press, etc.
I promised that we will never be like Google, which is near impossible to find a contact form, email or phone number.
#2 - Have an easy contact form on the Web Site
This is a big mistake that some smaller companies make. Big companies usually prefer to have a contact form (and sometimes no email address, like Comcast), but smaller companies don't add a form to their site because they think if they are providing an email address there is no need. Yes, it is better, but not 100% of the time.
Imagine that You're on an Internet terminal in an Airport; you've just found a bug and want to report it or you want to ask a question. For sure, that terminal won't have your email configured. So, if Sampa doesn't have a form to submit a question, the only way to do that is by finding the email address on the Sampa website, copying that address, going to the address bar of the browser, sign in to your email, create a new message, paste that email address and enter your question. Easy? Not so much.
#3 - Have a one-click easy contact form on the Design Site
This is one of the best features of Sampa. We have a persistent button on the top of the screen with a Ladybug () graphic called "Quick Feedback Tool". Click on that button and you've get a *one* field form to enter whatever you want to say to us: feedback, comment, question, praise, bug report, whatever. Click 'Send' and we get that message. There are many important aspects of this way of sending feedback:
It is always just 1-click away from anything that you can be doing on Sampa;
It has only 1 field -- free-form text -- you enter whatever you want.
It doesn't take you away from the current context. It opens a dialog box and once you send it you are back where you were before.
We send all the necessary technical information on the background to help us investigate the issue (which page you were in, what is your email address, site name, etc.), so we don't have to ask you for those things.
More than 90% of all feedback that we receive on Sampa comes from that simple form. About 5% comes from direct email contact, and about another 5% comes from the Contact Form on the Sampa.com website.
#4 - Send anniversary reminders
I don't like the term "Listen to your customer" (hence the title of this post). You should be proactive and engage them. Besides sending confirmation email when you create an account, we make sure that we send you a reminder for the 1 week and 1 month anniversary of your site. Both emails *proactively* ask for feedback and contain a link to a survey. Very few people fill that survey, but still, if we can get a couple of surveys filled per month is awesome.
#5 - Have a blog for the customers!
Heck, even Dell has a blog. When I say have a blog for the customer, I mean for your company to have a blog for the customer. Got it? If not, this is not a pre-sales, this is not a corporate blog, this is not a blog to talk about the challenges of building the product, or dealing with suppliers, or bragging, or cursing. It is a blog for the customer. The only things that should go into that blog is posts that will help your customer make better use of your product. Period. Now, at Sampa we don't have that (gosh, I'll have to fix that). What we have is two blogs in one: Customer Blog and Corporate Announcements blog.
#6 - Have a newsletter for the customers!
That is important and most Web 2.0 companies don't do that. They think it is annoying (because of annoying companies like Dell) and the truth is that it doesn't have to be annoying for the customer. First of all, the frequency has to be reasonable. Daily or weekly emails is a no-no in my book. Imagine if every company that I bought something from or I'm using one of their products would send me a weekly email. Do either monthly or bi-monthly emails.
The second important aspect is the content. Don't make it sound like a car dealer sales event. Make it useful. At Sampa we only announce the new things that will make customer enjoy more or learn more about our product. It always start with the new Customer features (we don't waste your time with our "better datacenter", "better backup", yada, yada) then we always explain how to get started with them. Sometimes, we might add a link to a survey (same survey as #4), and that is it.
#7 - Promptly reply technical questions.
Did you ever send a question to a company, and 5 days later you get an answer? And, you had already forgotten about that? That is bad. Some companies setup automated answers to acknowledge that they received your question. I find that ok, but not very useful. Bloglines does that, and on their case I find it annoying. My Datacenter (Isomedia) does that as well, and on that case I find it very useful. The difference: Bloglines is not a life-or-death situation, Isomedia is a life-or-death situation for my company.
Being realistic, Sampa is not a life-or-death product either, so we don't send you an automated "thank-you-for-contacting-us-here-is-your-ticket-number" email. On the other hand, no email or contact to Sampa goes unanswered. During business hours, I personally reply to every customer question in less than 30 minutes after receiving it. While I'm at home or on the go, I have my Smartphone (and sometimes my laptop) with me. If I can answer the question without an investigation I reply immediately, otherwise I can only answer when I'm back at the office. On the worst case scenario, a customer answer will take between 12-18 hours (yes, I work every day of the week).
#8 - Send survey to users.
I talked about that on #4 and #6. Before a "do", here is a "don't": Don't do that stupid in-page survey popup (like Microsoft and Seattle P-I). When I'm on their website I'm busy. I'm reading something or looking for something. I'll never stop my current task to answer a survey (unless I'm really mad and want to trash-talk them).
Email surveys on the other hand I do answer from time to time. Why? Because if I receive a link on an email for a survey I don't have to answer it immeditely. I can leave at my inbox and later, when I'm not so busy I can click on it and answer it. I also love when they state that average time it will take to answer a survey. From 5 to 10 minutes I'm ok. Anything above 15 minutes (or 10 questions) I'm out.
We use SurveyMonkey at Sampa (it is easy and for the simple stuff is free). Besides sending surveys on Anniversary and Newsletter emails, we also send "focused" surveys to get a specific kind of feedback. We try not to annoy too much our customers, but with some of them we treat each other on a first-name basis already.
#9 - Direct contact top users.
A top user is not only your more consistent user, he is also, potentially, your best advocate. He is a top user because what you are providing works really well for him/her. Do you know why? Do you know what they want to do? How they want to use the product? What they thought was the best and worst of the product?
We do contact our top users directly. By 'directly' I mean is not automated. We send them a personal email. We thank them for using the product, ask them what they are thinking, what they are trying to do, what they found hard to use, and, sometimes, we even ask if we can call them to have a 1:1 talk. About 80% of top users don't answer, or answer saying they are too busy and don't want to talk or discuss the product. The other 20% are ecstatic that somebody is asking them what they think.
#10 - Always leave a door open for feedback.
Every moment that we "touch" a customer, from when they sign in to when they receive a confirmation email, when we send them a survey or the monthly newsletter, every single time we tell them that any and all feedback is welcomed, positive or negative. Every time we answer a support question, we tell them to contact them at their convenience if we didn't solve the problem satisfactory. But listen carefully, this is not an impersonal corporate-talk asking for feedback. This is trully sincere and personal. I hate those taglines of customer support people "My goal is to make you enjoy more our product." or "We love that you're using products from XYZ Corp. Contact us at any time if you have problems".
That is probably the longest post I ever made, but it is worth it so it will keep me honest as we grow to be a big corporation. Here is caveat... All that I talked above is how to engage and get customer feedback. What I didn't talk about, yet, is how to use that customer feedback and this is one of the biggest traps that I know in business, and that is another post.
My favorite blog to customer focus is Creating Passionate Users. Kathy Sierra knows her stuff. Subscribe immediately.
Another talk at the Entrepreneur University was done by David Weld of SeaTab software. It was the worst talk of the day for me.
His goal was to talk about listening to your customer. He even brought a customer. It was lame. Why? Because if you listen to 1 of your customers and 1 of your customer thinks that you listen to him is not in any way either a proof that you listen to your customer or that you have mechanism in places to listen to your customer.
I trully, deeply believe in connecting with customers. This is more than listening to them, this is about having a dialog on many aspects of what you're providing.
Yesterday I wrote aboutAndy Sack turning off comments on his blog, actually, making it so that users need to sign in (which is as effective as turning off comments).
I assumed he got some hate comments and that is why he did. Which was quite possibly the case. Today he posted saying it was because of comment spam, which I tend to believe to be true.
I had a Typepad blog and (argh, we shouldn't talk the truth about our competitors openly... but I was a customer) Typepad does a very poor job helping customers fight comment spam on your blog.
Sampa has CAPTCHA (we prefer to call it "Word Verification") since Alpha 2 (we didn't have a blog on Alpha 1). Sampa also implements some neat black/gray list techniques that are very popular on search engines to prevent site spams (remember I worked on MSN Search?). On top of that, we have about 3 other "secrets" to reduce Spam to close to zero on comments. But enough about Sampa...
Andy, I started this discussion because yesterday I was going to reply to one of your posts with a suggestion. Now you lost that opportunity. You loose opportunities every time you post something on your blog because nobody very few people will be willing to sign in.
At minimum, enable Word Verification (CAPTCHA) and disable trackbacks. That will get rid of 99.99% of the Spam because spammers need efficiency, and Word Verification, althought can be circumvented, increases the price of posting a comment into your blog from 1/1,000,000 of a cent of a dollar to 1/1,000 of a cent of a dollar making it not worth for spammers anymore.
Example 3: You come home after work, doesn't say a word to your wife and expect diner to be served and the house to be clean.
Example 4: Despite the fact there are hundreds (maybe thousands) doing the same thing that you are doing, you demand to appear on some high traffic blog.
Thankfully, I'm none of those anymore (I was before). I prefer to assume that I'm not entitled to anything on this life and I must earn every single thing. If you do that, you'll see how different your life is going to feel.
This was scheduled to happen. I even wrote before about how AJAX can be misused.
I just read on John Cook's blog about this new product called GeoJoey. It is another one of those sites with map mashups, where you can write about your experience at some place.
I went to check the service and I was a bit surprise by the size of the URL on John Cook's blog. Then I notice that they are using 100% AJAX (translation: the site has only 1 page), and every parameter is being passed as a URL Hash string.
This is bad. This is beyond bad. This is super-duper bad.
Here are a few reasons why...
1) Search Engines
Search Engines are already not a big fan of query string (like in http://domain.com?id=123). Hash strings is simply absurd, like http://domain.com?id=123#hash. I doubet that any search engine uses that information. I'm actually willing to bet that no major search engine considers "#123" different from "#1234".
2) Email
You know the biggest enabler to make a product viral? Make so that people easily share the URL by email, like in the case of YouTube. The problem with huge URLs is that a lot of encoding, decoding, escaping, unescaping, and other nasty stuff happens between a sender and a receiver of an email, and it is not that hard to have URLs being broken between multiple lines (usually anything with more than 70 characters has the potential to be broken).
The URL that I'm seeing on my browser right now has 190 characters.
3) Forms
Any decent developer adds limitations to the input parameters that they can receive. Yes, the standard says that URLs can be 2048 characters. Actually, the standard doesn't really have a limit, but that is what a lot of people has been using. Turns out that a lot of sites and applications sets the limit at a way smaller size, like 128 characters, sometimes even 64 characters. Well, you won't be able to paste a GeoJoey link.
4) Usability
AJAX is great if used properly. AJAX is horrible in terms of users experience on the Web if over-used. Users, for the last 15 years or so, have been used to non-AJAX apps, with forms that load new pages, with clicks that load new pages. I'm not going to say that no page loads is better or worse, it is just different. When you try to push the envelope in creating a site that is 100% AJAX you pushed it too far. Users are not used to that, it will fell weird for most users.
Now, all that I said above is fact-based (challenge me in the comments if you believe I'm incorrect), for my personal opinion... This is just another map mashup site.
His list doesn’t come even close with the worst lies of engineers. It looks like he never talked directly to an engineer, only to the engineer’s manager that passed the message to him.
Here is what I think are the worst ten lies:
We are on track to ship on the scheduled date. (Read: we will be 2-3x late)
This feature will only take a week to add to the product. (read: the feature will take 3 weeks).
We should rewrite this component because it’s full of bugs. (Read: I can’t understand what the previous engineer did, so I think it is easy to write from scratch. It will have the same amount of bugs, but at least I’ll understand)
I’ve tested the feature and it’s working flawlessly. (Read: It’s working flawlessly for my way of doing things. Don’t try to use the drag-and-drop feature ‘cause that’s not how I use it)
This feature is easy to discovery (Read: After you read the documentation on page 37, third paragraph and enabled the checkbox on the options dialog the button will appear on the screen.)
This feature is easy to use. (Read: It’s easy to use as long as you understand relational databases, and/or object hierarchies, and/or state machines, ...)
Even my mother can use it. (Read: My mother has a degree in Computer Science and she develops her own apps, and she will find the feature easy to use.) Not a lie, but a misleading statement.
This feature is Pri 1 (actually, Pri 0). Everybody is going to love it. (Read: This feature is so cool to implement.)
The code is well documented. (Read: It’s documented in C++. Why? Don’t you speak C++?)
We should use inheritance, componentization and some design patterns to simplify the development. (Read: This is how we learnt in college and on many UML books. This is the only way to do it, no matter that this is mostly theoretical and very few people successfully accomplished “the perfect code”)
I could continue on and on with this list. I’ve worked on Microsoft for too many years with many different types of engineers to see all the lies. And don’t take me wrong, most engineers don’t even know they are lying, they truly believe when they say the things that they say.
We been having this debate for a long, long time on Sampa. What is an 'Active User'? Or, is 'Active User' the right way to track activity on the site.
Sampa is different from MySpace, for example. On MySpace an user is somebody that sign up to MySpace and created their page, so, they can say that each MySpace User is an Active User, otherwise they have to cancel their account.
Sampa is not a walled garden, and the sites created by our users are open to the Internet, like on Blogger. We could define 'Active User' by users that change their content on their web site on the last X days. We even prefer 'Active Sites' since you can have multiple editors and administrators of a single site.
Now, let's say that a site has not been updated for 6 months. The easy answer on this case is to consider that site not active anymore, but, what if the site does get a lot of traffic, even with the stale content? We called the people that sign up for Sampa 'Users' and the people that visits their websites 'Visitors'.
Technorati considers an Active Blog any blog that has been updated on the last 3 months. For a pure blog that makes sense, for a Blog like Sampa where you can have photo albums, Flickr/YouTube integration, pages, and other content, you might not post for a long time, but you still have your site.
Right now we have three definitions at Sampa:
'Active Sites': Sites that have high traffic in the last X days.
'Active Updated Sites': Sites that have been updated in the last W days.
'Active Users': Unique Users that visited Sampa Design Site on the last Y days.
'Active Visitors': Unique Users that visited any Sampa sites on the last Z days. (this is the typical way websites track their growth).
The question remains, how much is X, W, Y and Z? 30 days? 60 days? 90 days?
At the end of the day, we should have a single number. We still debating if the 'Sampa Activity' number should be a combo of the values above, or just the number of Unique Users on the public Sampa sites, which is more in-line with today's standard way of tracking usage of a site.
You know when "they" say that a consumer will tell on average 3 other people if he had a good experience with a product or company, but they will tell 6 other people if they had a bad experience?
Well, on the blogosphere seems to be more extreme than that. You'll see 10 posts trashing a product/company from an average blogger for each 1 post complimenting a product/company.
This is one of those rare "this-product-is-great" post...
Two weeks I need a new batch of business cards. The last time (which was the first time as well) I went to Kinko's. Horibble quality. I could see the pixels on the card! The color was awful and the paper not that great either.
This time I decided to find a more professional service, so I went online and investigated about 4 companies. After much debate (with myself), I decided to try Digital Room. A-mazing!
The quality of the card -- which is what matters the most for me -- is outstanding. It feels great. When I compare with business cards from other VCs, CEOs and Consultants, this is much better. The colors are outstanding.
Now, the big surprise, it cost just about $35 bucks for 1000 cards!
I strongly recommend them. I did a tricky design, with bleeding and they executed flawlessly. The bad thing is that the minimum order is 1000 cards, and that is just too much. I never did more than 250 at a time.
I was a really disappointed with the color calibration of my monitor, the card looked darker than I wanted, and the fonts end up a bit too big for me. Now I'll need to pass around 1,000 cards to need to order more cards. :)
On the footsteps of Evan Williams (Odeo) and Andy Sack (Judy's Book) talking about decisions, strategy and executions that didn't go so well on their company, I also decided to write my own version of it with regards to Sampa.
There are two reasons that I'm doing this. First, it helps me flush out some of the stuff on my head and think more clearly. Second, and most importantly, it can help some other entrepreneur out there that might look at this and learn about it.
The list is not in priority order, but more or less in occurrence order.
#1: Taking too long to get started
I had this idea 3 years before I left Microsoft and started Sampa, circa 2001. The idea was not very solid, but by 2003 it was mostly what you see today. Between 2003 and end of 2004 I tried to convince Microsoft on doing it. That was a huge waste of time. I should have packed my stuff and started Sampa back then.
#2: Being in Stealth-mode for too long
If you are a Startup that is entering a big market, there is no reason on being stealth. What? Do you think Microsoft or Google will look at your 2-person company and say let's copy what they are doing? Get real. Only about 5 people knew what I was doing for the first 9 months. After I started telling people, a wonderful thing happened. People started connecting me with other people doing similar or complementary work. If I could roll back time, I would have told everybody that I knew about what Sampa was going to be.
#3: Not having a partner in crime.
This has been said already on many books and blog posts. You are much more likely to succeed if you have somebody else working with you. Back when I started Sampa, it never occurred to me that somebody else would even be interested in joining me. This goes back to #2. If I had been more open about it, maybe I could have found somebody that had the same passions and motivations that I did.
#4: Taking too long to decide between Small Business vs. Consumer
Sampa was never meant to serve a medium or large-scale site. So, the obvious choices were Small Business or Consumer. I did what every ambitious (and dumb) people would do: I decided to go for both! This was mostly a R&D mistake at the time because I had no product, brand or marketing, it wouldn't affect those disciplines. But I spent a lot, a lot of time developing features for Small Businesses that are not useless. We actually even removed some of them from the site. Sampa is officially a consumer product.
#5: Not preserving user data during Alpha tests
Alpha was meant to be alpha, so, I didn't want to have to be backward compatible. Every Alpha release meant all user data was gone. Everybody had to sign up and try again. The vast majority of users understood that and were ok with it, except they didn't come back later.
#6: Swing too much vertical
Sampa was a very horizontal offering, meaning, it was a product without a niche. It works for anyone. Marketing 'experts' will tell you this is a bad strategy for Startups because you cannot get past the noise level and you never create a brand. The suggested solution is to go vertical, find a niche where your product can provide a great solution for their problem, build a few features focused on them and market like hell for that group. So, we picked a niche, then we started picking sub-niches, and spent a few months working only on features, branding and marketing strategy for those. That was too vertical. We are a horizontal platform, and this is how the service was designed to be. Being "too" vertical meant going against the grain of the platform. Right now we think we can pick a niche without being vertical.
#7: Hiring 'A' players that didn't fit the culture
You know how every single Entrepreneur book tells you to only hire 'A' players, the best -- well, they are wrong. You should hire the people that will produce the most results for your company, and that is usually a combination of excellent people and synergy with the existing culture. And culture being the broad term that represents everything a company is and has.
Every time I need to contact Google for whatever reason I know I won't find an email address, I might or might not find a form, and it will be a long and painful search for that info (ironic that Google has a mission of organizing all the information on the world).
Somebody signed up to Orkut using my email address. I received a link to confirm the email address, once I clicked on the link on the hope that I could say "Cancel this", I just got a message saying "Congratulations, your account has been activated."... What? No, no!
Ok, that is easy, I have the email, go there and click to delete the account. Well, if it was that simple. Orkut has a clearly marked link to delete account, I click on it and I can't do it because I don't have the original Orkut Username and Password. I need them to link them to my Google Account and then delete the Orkut profile.
I do have a Google Account which I can sign in to Orkut, but I can't cancel the Orkut profile because it was created with a Username and Password.
Microsoft Office just RTMed this week, which means in a couple of months people will be buying them on stores and it will come installed on their new PC.
The great news is that Outlook 2007 has embedded support for RSS. This is a super easy way for people to subscribe to their favorite blog/site without knowing a thing about feeds. Great news.
The bad news for blog owners, developers and marketing people is that you won't know when a user visiting your site to get a feed is using Outlook 2007. That sucks!
I found this post from Michael Affronti, a PM on the Outlook team, who gives some lame excuse why they cannot change the "user agent" string for those requests.
It is not always that you find a major website built by serious developers exposing an ugly call stack to their users. But AOL did it for me:
javax.servlet.ServletException: TEA: Length of String s is not a multiple of 8. com.aol.search.mvc.DecryptQueryServletFilter.doFilter(DecryptQueryServletFilter.java:112) com.aol.search.mvc.TestbedServletFilter.doFilter(TestbedServletFilter.java:104) com.aol.search.mvc.UserAgentBlockFilter.doFilter(UserAgentBlockFilter.java:218) com.aol.search.msrp.filters.RequestIDOverrideFilter.doFilterInternal(RequestIDOverrideFilter.java:82) com.aol.search.gsp.filters.AbstractConfiguredServletFilter.doFilter(AbstractConfiguredServletFilter.java:160) com.aol.search.mvc.LoggingServletFilterBase.doFilter(LoggingServletFilterBase.java:94) com.aol.search.gsp.filters.LogonTimestampServletFilter.doFilter(LogonTimestampServletFilter.java:94) com.aol.search.mvc.UserInfoRedirectFilter.doFilter(UserInfoRedirectFilter.java:242)
All that I did was to change the "encquery" parameter to a normal string. I wanted to see what type of encoding they were using. Apparently the do some cryptography because the function was called "DecryptQueryServletFilter". Why the heck would they cryptograph this value?
How are websites owners supposed to know what people are looking for on their site if the referer is an encrypted string. If Google had done that from day 1, AdSense and SEO would not exist today.
Subscribe to this Site
there was this big push back in 2002/2003 for us to be Customer Focused. It came from the top (Steve Ballmer), so it must be done. Back then, I thought I understood what that meant. I've read customer feedback, I participated on Usability Studies sessions, I pushed my developers to do the same, etc. So I must have been Customer Focused... 
) graphic called "Quick Feedback Tool". Click on that button and you've get a *one* field form to enter whatever you want to say to us: feedback, comment, question, praise, bug report, whatever. Click 'Send' and we get that message. There are many important aspects of this way of sending feedback:
September, 2007