Everything I've learned about podcasting over the last four years

An exhaustive how-to guide on getting starting thinking about and recording, editing, and publishing your first podcast.

The other day I realized that I'd been podcasting for over four years over at MetaFilter. It's a casual thing I and two other moderators do about once a month, but in that time I've amassed a great deal of knowledge mostly through trial and error. I've shared lots of these tips to friends before so I thought it might be handy to write down everything I've learned about podcasting and put it all in one place.

The Kevin Pollak Chat Show goes 2hrs+, and goes very deep into each interviewee's life.

Time for podcasting

Time commitment is the first stumbling block of podcasting. Podcasting is a special thing in that to me it feels perfectly natural to go on way longer than a typical radio or TV interview, and I am drawn to podcasts that are in the one to two hour range because the host and guests can cover a significant amount of ground and go very deep in interviews about a subject.

Chances are that you are busy doing other things so you'll need to make time for recording and more importantly, editing your podcast. I've found that most 2-3 person podcasts can easily go 60-90 minutes naturally if you are good friends and used to talking to one another. Anything less than 30 minutes of recording feels too short to me and I've often ended podcast interviews as a guest wondering if we had enough time to explain or go deep enough into anything.

I typically carve out two hours to do my podcast and I usually have to warn my two other hosts about 3 days in advance via email and once we coordinate schedules, we all show up at the agreed upon time ready to record. Editing is more important and I've found in my experience that there is an inverse relationship between length of podcast and hours of editing. My first couple podcast episodes were tight 20 minute long affairs. I was recording for about 45-60min at first so the editing was brutal, taking out lots of material, trimming conversations of all tangents, then I would spend hours cutting out ums, ahs, and pauses. My first 20min episode took about eight to ten hours to edit and nearly sapped all my motivation to ever do another podcast.

When Leo Laporte hits "record" it's all business and everything gets uploaded as the show. Efficient!

At some point in the last four years I was once a guest on a Leo Laporte TWiT-world show. The guy is an old radio pro and basically when he says go, the show is on, and everything that gets recorded soon goes online. I really envied the efficient, live style of his shows and after that I urged my other hosts to keep our meta-commentary to a minimum (we used to often say "I know you'll cut this but..." and talk for five minutes). At this point, we record for about 90 to 105 minutes and typically shows are around 80-90min long edited. Editing only takes me about two hours per show these days, both because the hosts and I all stay on topic and editing is streamlined to a single pass where I just cut out pauses and anything obvious, add music, and finalize it with podcast file info.

I would stress to anyone planning a podcast to give yourself enough time to get deep into the subject matter with your hosts and any possible guests and to also try and keep it as all-business as possible to minimize your editing time. Editing is the hardest part and a total time sink, but if you do your best upfront, you can minimize the toil of editing.


I listen to a lot of podcasts and the most typical format is 2 or 3 hosts and sometimes one guest. I've never subscribed to a single-person podcast before because I've yet to find a single-person-talking podcast that is interesting enough to stick with. I've tried a few but one person talking is usually pretty boring after a while. Two or three people chattering to each other is the most common format but it's possible to take it too far. The other day I heard a six person podcast that was an utter nightmare. Everyone talked over each other and when it came time to make small quips and jokes you had to wait for five people to make a wise crack. Stick to 2-3 people on your show.

Interviews are an obvious fit for podcasts and even if you have a good established show about a topic with a couple of hosts, bringing in the occasional guest is always a nice change. A pure interview show also works, though I'd say the 80 or so years of radio interview history may fight against you in terms of finding something new to do with an interview format. Still, if you can book a steady stream of guests and you're ok with scheduling time around your guests (they are after all, being on your show for free) then regular interviews can be great.

Keep calm and carry on podcasting

Before we dive into hardware and software and the nitty gritty details of recordings, it's important to think about your mood or your presence or whatever you want to call how you feel before you hit the record button. Remember to relax, even though you might be nervous at first. You're just a person sitting in a room in front of a computer talking, so try to avoid thinking about how famous your guest is or how anxious you are about not knowing what to say and be as relaxed as possible when you start. It's ok to do several takes if you're recording an introduction. On the flip side, it's good to relax but you don't want to come off as low-energy in the recording because that can quickly bore listeners. Try to let your enthusiasm for whatever subject the podcast is about come through. Podcasting is like any acquired skill: practice makes perfect, and now that I've done sixty shows and probably well over 100 hours of recording I'm completely calm when we start but it took 5-10 shows before that became the norm.


I know a lot of podcast hosts were former musicians and can talk for hours about microphone selection, positioning, and use, but I've tried some high end equipment as well as low-end stuff and my advice to anyone new to podcasting would be to stick with what is most simple and simply works. I would highly recommend a USB headset headphone/mic combo and I would say you should expect to pay about $30-50 for a decent headset. They are available on Amazon (here's a search) and at office stores these days and I've personally had great experiences with whatever Logitech is selling for about forty bucks.

Again, you can spend hundreds of dollars on microphones, splitters, A/V switch boxes and more, but in all honesty when recordings get mixed, edited, and compressed down to a mp3 in the end, a forty dollar headset sounds just fine.

If you're doing interviews with remote guests, buying an extra USB headset is a good idea as well. I know some shows will ship out a $40 USB headset/mic to guests a few days in advance to ensure they get a decent enough recording.


I use nothing but Macs, so this will be Mac-centric, but I typically record podcasts remotely through Skype, while each person in the podcast records locally using Call Recorder, a companion recorder app for Skype. In the radio business I hear this is called a "double-ender" but basically you record your conversations locally, then put the locally recorded voices back together in your audio editing app and the effect is close to sounding like you were in the same room together (you don't end up using the recorded remote skype sides of the conversation so voice quality doesn't really matter).

You start off installing Skype and Call Recorder (I set my Call Recorder settings to record in AAC at a quality setting of high), and every time you launch Skype a record window will automatically pop up. Now, it's not totally important for all your hosts to coordinate the exact moment when you start recording (just make sure everyone is recording before you start the show), but it's super important for the person hosting the Skype chat/call to keep everyone recording until they kill the call at the end. This creates a nice end point that you can later match up all the audio files to in your editor.

So you hit record on Call Recorder, run your show in Skype then have the host kill the call so everyone ends at the exact same moment. At this point, every person that recorded the show will have the show sitting in the default "Saved Calls" directory. You should also have a directory on your Mac called "Movie Tools" that installs with Call Recorder. You'll want to drag your recorded call over the app called "Split Movie Tracks" (or Split Sides of a Conversation). You'll get two tracks as a result, with the same file name as the original but with track 1 and track 2 appended to them. You'll just be using track 1. If you are doing the final editing, have your hosts/guests FTP their own split track 1 files to you (everyone usually uploads to their blog's web server in my case). Then you download the files and now you have everyone's raw locally recorded full-fidelity recordings.

I use Garageband for audio editing because it's pretty simple to use. Note there are some reported iChat/Garageband features specific to recording live podcasts but I've had Garageband crash on me multiple times while recording a podcast conversation, essentially losing hours of work in the process so I would never suggest anyone try recording live in Garageband.

Killing reverb in Garageband is step one.

You'll want to first set up your tracks in Garageband. I have two male hosts and one female host so I add new tracks to match. I immediately go into advanced audio properties (audio/edit on the right side panel) to set Master Reverb to 0. I don't know why Garageband insists on adding a crappy-sounding echo to every podcast voice but I kill it immediately. Next, I make sure I have one track for audio jingles (we have a theme song and sometimes toss in music or sound clips) and the main podcast track is for album art.

Setting up tracks at the start of the show in Garageband.

Once tracks are set up, I take each person's track 1 recording and drag it over the appropriate track. I line up the ends of the tracks to the same moment, since I made sure to close the call and cut everyone's recording at the same time. I add theme music at the top, then cut the beginning of the sound recordings to the point where we started the show.

Cutting across all three hosts' tracks in Garageband.

The specific steps are: move the play head to the right spot, highlight your voice tracks, hit command-t to cut across all three, then select the three tracks to the left and delete by hitting command-x to cut, Select your three remaining tracks together and slide them into the start of the podcast just after your theme song.

Sometimes, it can take Garageband a while to analye and show the audio spectrum for all your tracks, and there are times when I've first important tracks, saved the podcast in Garageband, and then left it open while I went to lunch only to return after to see nice peaks and waves in every track. This helps visually with editing, as you can learn to spot dead sound areas or high spots of clipping.

Garageband is pretty easy to edit audio in, but when you're dealing with a podcast, make sure any cuts you make are selected across all your tracks at once (in other words, if you want to cut out a 10 second pause where no one said anything, make sure you cut ten seconds from all tracks, not just one). Since you have multiple tracks covering a long time period it's important to move all your voice tracks together when editing so they never get out of sync with one another. Every time I mention "editing" when talking about a podcast, I just mean doing this over and over again, hundreds of times.

Try to conduct all your editing from start to finish, as adding a song to the beginning when all editing is complete is actually kind of difficult, requiring you to move possibly up to hundreds of track edits to make room for the song.

I go through and edit usually in just one pass though it might take you a few passes on your first show. You'll want to remove any pause that lasts more than a couple seconds, if there is a sneeze or lip smack that sounds annoying, you can highlight a single track and cut just those bits. When you are done, make sure the sound levels sound good and even across all your hosts and music tracks. This can be difficult and requires some trial and error. Lots of podcasts come out sounding too quiet but it is easy enough to solve (be sure to select a track and adjust the entire track's volume individually).

Updating a podcast track's info in iTunes before uploading.

I finish a track by using the Share menu to send the podcast to iTunes. This takes a few minutes to compress, convert, and import to iTunes. Once in iTunes, I've found that Garageband typically strips all my podcast info from the track so I have to Get Info on the track and provide the correct info. In the Info tab, set the genre to podcast and make sure the title and artists reflect your podcast. In the Options tab, check the box marked Remember Playback Position and you might want to tweak the volume up from None if it sounds quiet. On the lyrics tab I usually pop in my Podcast site URL, and on the Artwork tab I drag in a 300x300 podcast logo I've saved in GIF format (JPEG and PNG work too).

If you continue to have audio issues with the final file (like everyone is too quiet or one speaker is much louder than the rest) then you might want to try running your output mp3 file through the free program Levelator. It does a pretty amazing and fantastic job bringing everyone's levels up to match. The final file will sound a bit more compressed but your audience will appreciate being able to hear everyone for the loss in dynamic range.

Making show notes in TextMate to upload to the Podcast blog.

Since my podcast is a recap of a website and we're talking about other sites and online things we found on the podcast, I produce notes for each show with every URL mentioned so listeners at a computer can follow along and click on stuff as we are discussing them. This isn't quite a transcript, but it is nice, especially if you are watching a video on a podcast and talking about it and need a way for the listener to be able to watch that video too. I've heard there are resources out there that will do full text transcriptions at about $1/minute but I find the show notes are a pretty good substitute. Also, if you get really efficient at editing, doing up the HTML for your notes post on your podcast show blog is a good way to pass the time and have something that adds to the show at the same time.

Podcast Hosting

Now, chances are you have a blog somewhere and you might want to go to the trouble of doing a podcast-specific blog that downloads from your site, but I've found the more simple approach is to use a cheap podcast service that does all the hosting and podcast feed-making for you. Liberated Syndication (libsyn) is dead simple to use, and starts at just $5/month for a 50Mb podcasting plan. Soon after I started podcasting I moved up to their 250 plan for $15/month but keep in mind the 50Mb and 250Mb are meant to be your "active" file storage earmarks. The "active" window is the sum storage of your shows for the past month. You can have four years of archives like I do totalling many hundreds of Mb, but they only charge you for what is active, not in your archive. Additionally, they don't seem to monitor bandwidth, so if you have a popular show, you won't be docked for thousands of downloads. For five to fifteen bucks a month, it's a super reliable and easy to use service. I upload my files to Libsyn (via FTP, which they support), then I make a new podcast episode in their publishing platform. The libsyn publisher makes a nice podcast blog and podcast feed, but I use libsyn as background plumbing.

The next thing you'll want to do is create a new feed at Feedburner. Plop in your libsyn podcast feed, make sure you tell Feedburner it's a podcast feed and give it a handy easy to read name. You'll use your Feedburner feed to display on your blog as well as to apply to be included in iTunes' podcast listings (Open iTunes, go to the iTunes Store, select Podcasts, then look for the Submit a new Podcast option in the side menu). Sticking with a Feedburner feed means you can move your podcast to other systems someday if libsyn doesn't fit your needs.

Finally, you'll probably want a blog to post all your shows to. It helps to get a good domain and simply map the domain to whatever blog software you prefer. You can promote the podcast with links to it in iTunes and provide a link to your feedburner feed. Typically I make a new post for each episode of the show and I embed a player (libsyn provides an embeded player for each episode or the entire show's archives).

That's about it! Anyone can go from zero to podcasting for less than $100 and you'll definitely get better with practice, so jump in and give it a try. I've found that after a year of semi-regular recordings I not only got better at recording, editing, and talking in a relaxed way, it also helped my public speaking and definitely helped me be relaxed when doing radio and other podcast interviews.

Disclaimer in case it matters: I don't make any sort of kick back off any recommendation above, I honestly like libysn's hosting, logitech's headphones, and all the software mentioned.

How to pick an accountant for your online business

Finding an accountant that understands the internet isn't easy and after going through half a dozen myself, I came up with some tips and approaches for finding the right one.

Taxes are a necessary evil in the course of doing business, but core to financing a government and living in a society. While the internet offers tons of tax advice both good and bad, I feel finances are important enough to warrant bringing in outside professional help. If you're a geek, the temptation is to think "this is math — I know numbers!" but what you might be ignoring is the additional tax of having to learn an extremely complicated system. Leave it to the pros so you can focus on the thing you do uniquely well.

For me, having an accountant handle the complicated state of my finances also reduces the stress I get from using TurboTax each year and wondering if I did anything wrong. I want to state upfront that I'm not advocating cheating or cutting corners, but I've found in my own experience that the difference between a good accountant and a bad one can cost you thousands of dollars. A good accountant recognizes all the costs of running an online business, offers tips for good investments (that in turn, reduce taxes), and offers advice on how best to grow your business. Bad accountants miss out on all those things and simply give you a large bill due each Spring.

Compared to a traditional business, starting an internet business is trivial and carries a great number of advantages. You don't need a staff right away (as long as you do all the work). You don't need a storefront, inventory, or even anything to sell. You won't be limited by the hours of the day or your physical location. You don't need the permission of anyone to start. With just a single webpage you can start earning revenue through ads or other measures and begin from there building a business. I know people that have monthly expenses as low as $100 to rent a colocated server that brings in $10,000 per month. While that is amazing from the point of view of economies of scale, when it comes to taxes, it can be really problematic.

The rules governing taxes on business are geared towards the traditional, capital-intensive types of businesses (it always feels like half of the questions in TurboTax are about whether or not I own a farm). If you don't have to buy parts, pay a large staff, or purchase trucks to move products around, you don't have a lot of options when it comes to built-in deductions against your revenue. For the first few years I started making some money online, I only wrote off a minimum of computer equipment and a portion of my hosting and bandwidth fees. Every time I'd discuss taxes with friends, I'd hear new ideas for things I should have written off. After a few years of this, my finances and tax bills started to grow to the point where I felt a professional could help out.

Every Spring I'd take my 1099s to a new local accountant to run the numbers. What I found after using several local accountants was that they just didn't understand how the internet worked. They would ask me about equipment (minimal — just a laptop), how much I drove (none, I do it all online from my home office), and how many employees I had (zero, though I've paid programmers and moderators as contractors for the past couple years). When asked for ideas of how I could reduce my taxes, they were stumped. I realized I wasn't using the right accountant when I ran my taxes through TurboTax online and found I saved several hundred dollars over what my own accountant was telling me to pay. And that was before paying the accountant.

All about shades of gray

While specific tax rules are fairly black and white, when you look at a tax return and consider there are thousands of decisions based on individual tax rules, the end result of where you balance revenue and deductions starts to become more art than science. I realized that several of my past accountants gave conflicting information and most would take a conservative approach by default. Only one accountant mentioned deductions from retirement accounts. I would argue with most accountants over things like how much goofing off I do using my laptop or my cable modem, even though being a "professional blogger" kind of meant that any and all web surfing might end up as content on my sites.

Earlier this year, I stumbled upon this post from someone offering tax tips for creative professionals. While I didn't think I'd use most of the tips, it definitely was a step in the right direction and I ended up ordering the book by the author. The book is loaded cover to cover with creative ideas on how to bend tax rules to your advantage and even though in my opinion about 1/2 of them seemed really dodgy and probably left you open for an audit, the book served as a good reference point when talking to accountants. The ideas in the book were beyond my comfort zone of things I'd do with my own finances, but I wanted to find someone that was somewhere in the middle.

By the time I read the book, I was getting advice from a local and very conservative accountant. On a zero to ten scale of conservative-to-creative accounting attitudes, I'd place that accountant at a 1 or 2 while I'd put the advice in the book at a solid 10. As April 2007 began approaching, I realized I was looking for an accountant to handle my upcoming taxes and I needed to find someone closer to my line of thinking, maybe at a 7 on my internal scale. Aggressive, but still safely within generally accepted accounting principles.

Finding the right one

I've gone through five local accountants in three years by simply scouring the phonebook and talking to them. It's not the best way to go, but you can eventually find someone to work with this way. After three years of searching I found a great (non-local) accountant by asking around among friends that are also involved with internet businesses. I'd definitely suggest using your personal network of friends and colleagues as they're likely to have found a tax pro familiar with the industry.

If you have a lawyer you're happy with, ask them who their clients work with. Like in any social network, the people who are the best at what they do tend to be connected to other people who are also very good. Or at the very least they can let you know if they've heard bad things about a potential accountant.

Go with loaded questions

Interviewing accountants is pretty straightforward. Once you decide on a specific accountant, you schedule an hour with them and they'll ask you about your business and your finances. This will be the first indication of whether or not the accountant is familiar with online businesses. Then they'll let you ask them questions and you'll be able to hit them with anything on your mind. Usually they'll charge you for an hour of their time and it is typically around $200 or so.

When you ask questions, it's good to feel them out a bit by asking very specific questions using your actual revenue numbers. If you've read the book or blog post I mentioned earlier, you might want to ask an accountant about some of the ideas you are comfortable with, as most find them fairly unconventional. If you ask about something the accountant disagrees with or warns against doing to save taxes, make sure their explanation makes sense and their demeanor is something you can work with.

My best piece of advice is to ask a prospective accountant questions that you already know the answer to. If you build websites, chances are you can deduct at least some part of any device that displays HTML and that you use for testing your sites and apps. Ask an accountant if you can write off a new iPhone (hopefully they ask if it's your official business phone and how you'd use the device). If you run a topic-centric blog or write about a specific industry, ask them if you can write off purchases that you review extensively on the site. While the accountant might not give you an answer 100% in the affirmative, it's good to gauge their response against your personal thinking on the issue. You should also ask about specific deductions that apply to you. If you have kids, ask about childcare tax credits (I knew I qualified for this but one accountant missed it on my taxes last year) and be sure to ask about retirement accounts. For most small internet businesses, you can setup a SEP-IRA that deducts money you put towards retirement right off your total revenue. It's not a 1-to-1 deduction but you can reduce your total taxes while you plan for the future. Only two accountants I spoke with even mentioned IRAs when I first talked to them. Only one of those two explained each and every option before I had to ask.

Accountant Showdown 2007

By April of this year, I had a local accountant I was wary of, I had a big city accountant suggested by a friend, and I had my old familiar buddy, TurboTax Online. I figured my finances were worth getting a second opinion about, so I decided to spend the money to have my taxes done all three ways. My local accountant charged about $350, the big city one was about $550, while TurboTax's business level with all the options ran me $99.

In the end, the big city accountant asked me the right questions and figured out a couple deductions I didn't know I qualified for, and saved me $1500 below what TurboTax came up with. I probably could have gotten my TurboTax return to match but I probably mis-read one of the hundreds of questions lobbed at me during the online process. The local accountant I wasn't a fan of turned out to be the worst option, coming up $1200 over my own TurboTax return. I suspect she left off a few business expenses I listed.

What's the point of having an accountant, can't I just use TurboTax and Google?

After doing my taxes online for 8 years and going through six accountants in the past four years, I've come to a few conclusions:

  • TurboTax Online is pretty dang good and can get you 80% of the way

All but the best accountants I've worked with didn't do any better than me using TurboTax on my own.

  • If you have time to do research and read up on tax info, TurboTax + your knowledge is probably equal to most accountants

This is a big "if", but if you have the time and you won't be satisfied unless you handle it yourself, a few books and a lot of reading at MotleyFool.com can you pretty close to the best advice you'll find in the wild. Then again, if you're dealing with thousands of dollars and you're busy, it's best to try out an expert instead.

  • Quarterly tax payments are your friend

If you've ever run your own business or been a freelancer, you know the pain of scrambling to pay an unexpectedly large tax bill. Be aggressive with the amounts you pay on your quarterly payment schedule and you'll spare yourself a lot of pain down the line.

  • Even if you handle taxes yourself, you should see a trusted accountant every December with your estimated year-end totals on revenue and spending to figure out where you'll stand come April

Every tax day for the last four or five years I've owed the IRS thousands of dollars. For the last two years, I've met with an accountant in December to handicap where I'd stand come April. It's been a godsend because I've been prepared both times for the worst, and had plenty of savings ready to pay the bill. A December accountant visit is also good for determining if your estimated taxes are inline with your earnings -- whether you should be putting away more or less the same amount the next year. Since first quarter estimated taxes are also due around April 15th, it helps to know that ahead of time as well.

How Ads Really Work: Superfans and Noobs

The obvious way to make the most money possible is to plaster your sites in as many ads as your visitors can stand. But if you look at your traffic, chances are you'll lose little by turning ads off completely for your biggest fans.

Online advertising is hot right now (the big web properties have dropped billions of dollars on advertising companies in just the past few weeks) and I'm seeing all kinds of creative ways to stuff more and more advertising into sites. Giant flash ads that launch when you accidentally roll over them with your mouse, auto-playing video blaring at you while you're trying to read something, and those awful pop-up things like Snap previews and IntelliText ads. It's getting to the point where, frankly, things have gone a little nuts — these ads might be supporting a website financially, but they're hardly nurturing their communities. Though I've written about the wonders of Google's AdSense, I've also written about why advertising in RSS is a bad idea. Today I hope to convince a few publishers that there are times when turning ads off is a good idea.

The two types of visitors that matter

Last summer I started using Google Analytics to record behavior on my sites and it wasn't until several months later that I started delving into the statistics. My sites had lots of revisits from a core group of regular readers and then it would trail off to lower numbers looking at fewer pages. This was a totally expected trend and meshed with how I've maintained sites for years. The conventional wisdom is that you create a site you love and hope that others will find it and like it as well, and then you keep writing, photographing, designing and creating for that audience. Your biggest fans help drive the site and their enjoyment springs you out of bed every morning. A second priority, but still quite important, is trying to convert new visitors to fans. That's how I've developed, maintained, and thought about sites for years. But there was one important group I underestimated.

My reports from Google Analytics were the first time I noticed the huge numbers of people who were new to my sites. I expected it on one site that gets a lot of specific traffic from Google searches, but then I noticed it on other sites with completely different areas of focus. Then I started asking friends who run other sites and heard they had the same results. It didn't matter how big or small the site was, how narrow focused or completely open-ended the content was. The biggest single group of visitors to these sites were people who had never seen them before and would never return again. Among my informal polling of friends and my own sites, the lowest percentage of one-time visitors was 53%. Some sites had as much as 75% of their traffic come from people that had only visited once.


I'd long heard that traffic patterns followed a power law, with the most traffic from a small group of super-users, but I'm finding my audience falls into a bimodal distribution: tons of one-time visitors and the balance at the opposite end of the spectrum with lots of repeat visits, with small amounts in between. This is important because those two types of visitors ("Superfans" for the heavy repeat visitors, "Noobs" for the newbie one-time visitors) interact with your site in dramatically different ways.

How ads really work

When people use search engines to reach your site, they're looking for something. Chances are, your site probably doesn't have what they are looking for, but with context-sensitive ads like Google's Adsense, they'll likely see ads for things related to their search terms. Imagine that — ads that actually make a page more valuable to readers, not just the site owners. Random people searching for information are much more likely to click on those related text ads if the ads help them find what they are looking for. Compare that to a regular visitor that comes to your site dozens of times a week: How often are they going to click on any ads? How quickly will they learn to visually filter out the ads entirely from the experience? Superfans develop banner blindness extremely quickly.

What I realized when I looked at my Google Analytics reports was that the majority of ad clicks are coming from these one-time visitors looking for information. I do it myself when searching, especially if it's for a product of some type. I'll search, dive into the results, and if the top 5 don't have what I'm looking for, I'm very likely to click on related ads to see if that's what I'm looking for. New visitors to a site love to click on anything that brings them closer to their goal, and often times that's an ad. This, in essence, is the entire business model of per-click advertising.

Bringing it together

About five years ago I started toying with advertising on my sites, but I wasn't making much and at the same time I was reading one email after another complaining about the new ugly ads. When I was considering adding some graphical ads I asked a few friends what I could do to lessen the pain on my active membership. "Turn them off for members" was the suggestion from Rusty Foster. It was simple, yet brilliant, and I decided to give it a try.

My hunch was it wouldn't change the income very much and I'd get a lot of happier users in return. What I found was that overnight, the click-through rate increased. The overall impressions went down, but non-members were clicking much more than frequent visitors. For a pay-per-click model like Google Adsense, this meant I lost virtually nothing by turning ads off for the most frequent visitors, and the numbers increased for what did get served. Members of the site were much happier with the ad-free experience and it gave regular readers an incentive to become full-fledged members of the site. For the majority of total visitors to my sites, those one-time random search result drive-bys, nothing changed for the better or for the worse.

Since I started doing things this way on MetaFilter, a lot of other services and communities have done the same. Rusty did it with Kuro5hin after hearing how well it went for me. In the early days of Flickr, they used to show text ads alongside photos for non-logged in users. I notice that Vox inserts ads in between posts on my neighborhood page whenever my login cookie expires — logged in users see less intrusive ads. A few months ago, I was looking at a Twitter message I left and noticed there were these odd floating text ads near my icon, where my message was popping out. I sent off an email to Ev, condensing the gist of this entire essay into an email and mentioning how Twitter permalink pages were visually confusing — it looked like I was saying something about a text ad. Ev replied quickly, saying they were thinking about doing it and the next day ads were off when I was logged in.


There are of course some downsides to doing this. Some advertisers really want to reach your core membership, especially for real techno-geek type products. Also, if you're serving up ads on an impression basis (like CPM), you will of course lose out on a decent percentage of your traffic. On the other hand, if you consider that your biggest fans might load a few thousand pages per year on your site, chances are you'll just be losing a few potential bucks for each of them which should pale in comparison to the intangible value they bring in return.

The easiest way to implement this is on a site that requires registration — members don't see ads, people without a login cookie see them, but what about sites that don't have strict membership status? It's still possible to do something like this, by using some simple cookies. Set a cookie for new visitors and simply look for the existence of a cookie when serving ads. If you want to get fancy, you could even count the number of visits and clear any ads for people that visit more than x times a month. In a few minutes of googling, this intro to cookies looked like it covered pretty much all you'd need to know to implement this.


If you've got a site/service that is ad-supported but also has some membership, do your best users a favor and think about turning off ads (I'm looking at you Facebook and LinkedIn). It's likely to make a small dent in your bottom line but will pay off with a better user experience for your site's biggest fans.

(and yes, I should state upfront that I did indeed add advertising from The Deck to this site recently and I don't filter it out for repeat visitors. It's not Adsense and I actually like the products being advertised and it's done in a clean non-obtrusive way.)

Some Community Tips for 2007

Seven tips on how to run a successful community.

Every year or so I write a long post or do a presentation at a conference on the subject of community. Each time I approach the subject, I take what I've already written and add to it with recent things I've learned or learned long before and only recently realized. To prepare for an upcoming presentation, I decided to write down stuff I've learned/realized in the last 12 months. I suspect I'll be revisiting this topic many times on this blog but I wanted to kick off this first foray into community with a list of stuff I've been thinking about recently, but haven't written much about yet. If you want a list of introductory tips, one of my first essays was this one from 2001. Consider this more of a list of advanced tips.

Before we get started, first I want to rant a bit. I hate the term User Generated Content. I never use the phrase when talking about this stuff and I'll never use it when writing about it. I consider it a pejorative that reveals a lot about the person saying it. It makes members of your site feel like dutiful robots, crapping content that you convert into cash. The proper (respectful) term is community, and running one is a real challenge. If you're building a community you have to love what you're doing and be the best member of it. It takes great care and patience to create a space others will share and you have to nurture it and reward your best contributors. It's a decidedly human endeavor with few, if any, technical shortcuts. Ok, back to the tips.

1. Take emotion out of decisions

It goes without saying that if you want to run a successful community, you have to be extremely patient with people and give them second chances and the benefit of the doubt. It's also vital to the health of the community that you don't make rash decisions or even rational decisions that can appear as if you were acting on impulse. For communities run by just one person, this can be a challenge. No matter how patient you are, inevitably someday you'll encounter a user trolling the site for laughs, or someone copying your RSS and/or design, or making false claims on your own site, and your impulse will be to lash out at the user causing problems, to ban their account, and delete stuff they've written.

For the first 3 or 4 years of MetaFilter, I succumbed to emotion from time to time, usually late at night after a long day or first thing in the morning before I had fully woken up. I'd read something (or misread, as often is the case with simple text that doesn't convey the way in which something was meant) and take it personally or take it the wrong way and tear someone a new one. The usual result was me looking like an ass and a bunch of members disagreeing with me and a little while later I'd realize I messed up and apologize.

Once I got other moderators on board, this became much, much easier to do. Whenever something on the site irritated me, I now had someone to bounce it off of. Other moderators often hadn't seen whatever got under my skin and could often give an impartial judgment on whether or not anything should be done about it. If you don't have other moderators, it's a good idea to instead sit on decisions you make when you're feeling reactive. After a few years of making occasional mistakes, I'd take to waiting until morning if I saw something late at night that I didn't like. If it was first thing in the morning, I'd purposely put off the decision until lunchtime. In almost every case, by the time I revisited something, other members had dealt with it or I would realize I read something the wrong way at first.

2. Talk like a human, not a robot

If you elect to take your own personal emotions out of major moderation decisions, you of course run the risk of going too far. No one wants to follow a community run by a passionless robot.

"Be human" is popular piece of advice I'm reading about and hearing at conferences this year and I'll write more about the subject later on, but it's important that members of your community know the leaders are humans themselves. This advice can manifest in many ways:

  • When making moderation decisions, don't quote Terms of Service verbatim or legal code. Be honest and phrase things exactly like you would if you were talking to someone standing in front of you. "Hey man, we appreciate your enthusiasm, but what you're doing is uncool because..." is a lot better than "As per the user agreement, you must cease harrassment of other members per Section 10.9.a..."
  • Capture errors on your site and make them friendly to users instead of the default language-specific cryptic messages. Put your own email (or a developer's if you didn't program it) on error messages so people can tell you when they spot a bug and help you fix it (of course, you could also automatically send an error dump email to developers on the server side when errors are encountered, but visitors should also have the option of direct contact).
  • Be the best member of your site. Lead by example by participating as much as you can in your own community. This is a good way to attract other well-intentioned members of your site and also reminds everyone a real person is behind it all and building the best community they can for everyone. Speak honestly and be supportive of other members. When I think of all the communities I'm a part of, the ones I love are the ones I see the creators using everyday.

3. Give people something they can be proud of

If I had to give a reason why most newspaper blogs are filled with cranky screeds posted anonymously, I'd have to say having a generic blank comment form is key. Most every community that I contribute to offers a comprehensive user profile/history page, letting members customize to their hearts content and allow their profile to reflect their personality. When I think of mainstream news, TV, and newspaper sites trying to solicit comments from readers, I've yet to find something close to even a basic community site. The New York Times requires me to register to read most stories, but their blog system gives me a blank generic comment form when I want to comment on a blog post.

I'd love to see a large paper like the NYT implement a real community system. Based on my existing NYT login, I'd love if I had a profile page on their site, tied to any comment I left on a blog or any article I wrote for the paper (I'm sure I'm in the minority here, but there are writers for the NYT that would also be active on the site). Let me list my blog URL and track any posts I make about NYT articles on my profile page (the NYT already has a "most blogged" feature on their site). Feel free to show me ads that would actually make sense (example: I don't live in NYC, but I see NYC ads on the site -- you might want to pitch me home delivery or general ads aimed at out-of-towners) based on my profile.

If you gave readers a real profile page on a real community system at a newspaper site, I suspect the quality of contributions would go way up. Of course, you'd still get trolls and griefers trying to game the system, but the remainder of readers would post more often and post better things. Heck, you could even let readers connect to their friends that also read the site and offer tools useful to members (like "your friends liked the following articles") as well as gain additional traffic from repeat member visits.

4. Bring users in during community decisions

If the Digg HD-DVD encryption key fiasco taught us anything, it's that you can't make rash top-down decisions and expect your community to be okay with it. For MetaFilter, I run an entire forum devoted to discussing the site itself. I float new ideas and new UI enhancements there and anyone else can start a thread about some aspect of the site. When I have to make a tough decision, I mention it in a new thread and get the members' reactions and often tailor the final result based on their feedback. Like the Satisfaction blog described, if Kevin Rose posted something saying "hey we're kind of between a rock and hard place — we've been threatened with legal action over a post and would like to remove it to comply with the request" there'd still be a bit of an outcry (and that's why stuff like Chilling Effects exists), but it would have been nowhere near the reaction they got.

Welcoming the opinions of users gives the community owner(s) valuable feedback and gives members another way to contribute positively to the community.

5. Moderation is a full-time job

Don't underestimate the amount of work it takes to maintain, moderate, and keep tabs on a community. Often a single person can create, develop, and launch a new community site in just a couple months, and spend the better part of most days fixing bugs and creating new features. Once a community becomes a bustling place filled with thousands of users — if a single person is still running the community — chances are they don't have any time for writing code anymore.

If you've got an existing site/service that you're planning to add a community or social component to, don't expect someone with a full workload to simply take it on and spend a few minutes here and there maintaining it. Your best bet would be devoting someone full-time to the effort.

6. Metrics spread the work out

While moderation is a full-time job, it helps to make the job as easy as possible because moderators never have enough time to police everything. No one can read every single thing posted in every single place on most community websites. If you've ever used craigslist, you've probably seen those innocuous "flag this post" links sprinkled around the edges of postings. I took a page from Craigslist and implemented a simple user flagging system last year on MetaFilter. It's a basic mechanism that gives the community a policing outlet, but beyond the simple act of empowering users to help you moderate a large site, if you build the right toolset you can save yourself loads of time and stress moderating content.

My flagging system records the id of the item being flagged, the person flagging it, and the person that wrote the item. With this data I can create several useful views of the data. The easiest is to simply do a stream of recently flagged items, and use that as a gauge for stuff as it happens. More useful than that is to start grouping things based on total flag votes, then sort by most flagged over a short period of time. This gives a snapshot of recent hotspots and has proved the most useful tool of all. With the item author data, you can also compile lists of all-time problem users as well as most problematic users over the last month. I've found when looking at the past 30 days data, you can easily see if someone made just one or two mistakes in the community (if everyone flagged the same items they wrote) or if they're more of a chronic problem on the site (if lots of people flag dozens of their contributions). All of these tools are available in an admin dashboard-style interface.

On the flipside, I also implemented a favorites system throughout the site and do similar things to help publicize the best bits on the site (favorites are all public, flagging is not). Metrics are tremendously helpful tools that are pretty easy to implement. About the hardest part is figuring out what to do with the data, and writing the necessary SQL queries to get what you need.

7. Guidelines not rules

While it may seem like simple semantics, I personally shy away from trying to run a community based on hard and fast rules, and instead try to steer members into following community norms in looser guideline form. This often works for the majority of members that want a nice, respectful community. Once you start down the path of absolutes and rules you'll quickly end up in two bad places. 1) you'll get the edgecase loving lawyer/engineer types that will argue and interpret rules ad infinitum and break them just to see what happens. These people will drive you crazy. 2) you will find yourself in a situation where you have to make a bad decision you know is unfair, but you have to because it was one of The Rules That Got Broken. Guidelines allow for nuance and though it's hard to scale nuance in a large community environment, it's another way you can run a site like a human and not a lawyerbot.

I've been at both ends of this issue, as a user and a community creator, both with rules and guidelines, and I prefer the loosey goosey guideline approach. Write up a page of things you'd consider "ways to be a worthwhile member of this community" and "things you probably shouldn't do" and explain the approach when needed, but don't bother trying to come up with a hundred bullet points of things you can and can't do because once you go down the path of even a few rules, you'll soon find yourself at the top of a heap of laws that constantly have to be added to in order to please your most argumentative members, while at the same time having those rules hold you back and force you to bring the hammer down on people that accidentally crossed a line.

How to talk to the press

A short guide to what works and what doesn't when talking to reporters

Sometime around the year 2000 I started getting requests for interviews with reporters and it never really stopped. These days I talk to two or three newspaper reporters a week, amounting to hundreds of interactions with the press in the past few years. Based on what I've learned, here's a list of good and bad aspects, followed by my best and worst experiences, and some final thoughts.

Things good reporters do

This sounds simple, but few reporters do it well: know the subject. A good reporter does research on the interview subjects and knows the story they are covering forwards and backwards. The best interviews I've had were with people obviously familiar with my work. A great interview with a reporter feels like a natural conversation with a friend you've never met. They'll ask you questions that you haven't answered a hundred times before, and really dive into your experiences that led to something newsworthy. A good interview will feel open ended and go wherever the conversation leads. A good reporter will send you an email when the article is posted and thank you for your time. If you notice any of these qualities, relish the opportunity because these kinds of interviews account for maybe 5% of the interviews I've ever done. If you get a call from a fact-checker who verifies your quotes or information for an article, then go buy a lottery ticket, because you've just beaten incredible odds.

Things bad reporters do

The biggest sign you're in a bad interview is when you can hear a keyboard involved. Now, some good reporters might not be able to take notes on paper or remember the conversation afterwards and will ask for pauses to get some stuff down, but the worst reporters I've dealt with simply type everything they hear, as they hear it. The most astonishing thing is I'm often mis-quoted by one of these stenographer-interviews. You'd think they'd get the statements right if they were writing them down as they heard it. Bad reporters taking notes as they talk are often distracted because humans aren't that great at multi-tasking. They'll ask basic questions you've answered in other interviews. The interview will take twice as long as it should because you'll have to keep pausing while they type. It's a terrible, unnatural way to talk to someone and the resulting article is almost always lame.

Another bad sign is the reporter that calls and asks long detailed leading questions. When the interview comes out, you'll learn why the questions seemed so detailed: it was fill-in-the-blank reporting. Reporters that file stories like this will basically write an entire article and then at the end, call a couple people to fill in the blanks with quotes supporting points of view they already wrote about. They'll read a paragraph-long question to you and if you answer in a way that wasn't exactly what they want, they'll re-ask using the answer they wanted anyway (ex. reporter: "you said it doesn't seem that way? huh, you don't think it could all be a fake?" you: "I guess?"). When they get the slightest shred of a quote they wanted, that's all you'll see in the article the next day.

Lastly, I'd say watch out for reporters that delve deep and want answers to overly personal questions (keeping in mind we're talking about technology reporters that don't normally ask personal questions). This one took the longest for me to learn, but you don't have to answer any question you don't want to. Remember that being interviewed is entirely a volunteer effort on your part. Early on I had a couple tough interviews where reporters asked for how much money I was making, questions about my wife and family, and questions about where I lived. They'd remind me it was for the New York Times or Wall Street Journal and say their readers deserved to know, making me feel like the bad guy for not answering. As a meek computer dork, until recently I held onto a belief that every reporter I talked to was a respected, wise elder that was doing noble work. If you ever feel remotely uncomfortable answering a question, stop the reporter ASAP and say you won't answer the question and that they should move on. If you do say something you didn't want to, call back the reporter as soon as you can and tell them which question was off-limits and that they can't print it.

I remember complaining to a friend after such an incident, saying "Man, that reporter was brutal, we talked about all sorts of things I bet I'll regret later" to which he replied "Why did you say anything at all? He didn't have a gun to your head. In the future, don't say anything you wouldn't want to see plastered as the headline." It was blunt, but great advice, and something I've stuck to ever since.

Things all reporters do

Though it's tempting to see the press as either intimidating authorities or adversarial enemies, the truth is, you'll get the most value out of the press by having good relationships with the reporters that are most valuable for your industry or area of expertise.

One of the toughest things to keep in mind, especially if you haven't done a lot of press, is that reporters are just people doing their jobs. So that means they want to be good at it, please their bosses, and impress their peers. It also means they often want to get things done quickly or on time for their deadlines, they might be begrudgingly covering a story they were assigned that they're not that interested in, or they might be phoning it in while looking for another job.

If you want to increase the chances of good coverage, of getting a quote in a story, or having your work mentioned, one of the best things to do is simply make the reporter's job easier. Recommend other good sources of information (peers, people involved in the story, people that cover the subject), and have a portfolio or background info on yourself and your work available. A standard short biography and some quick links to information about your area of expertise are great for giving a writer the Cliff's Notes version of your life. If you're not familiar with their publication (and let's face it, most of the press you'll get, at least initially, is with papers that you might not read) take a quick glimpse and see what kind of pieces they usually do, and picture how you'll fit into the story. Google the reporter's name.

These days, a lot of reporters have blogs of their own. Skim through them, get a feel for what their personal obsessions or hobbies are, and drop in references to their own areas of interest. A genuine connection with a reporter doesn't just increase the chances you'll be happy with the story they write; It's also a great way to encourage them to come back to you for quotes or follow-ups in future stories.

On the record, off the record

We hear the terms "off the record" and "on the record" so often, it's easy to forget what they actually are supposed to mean. Though the specific definitions can vary, assume every single thing you say to a reporter is on the record and will appear in print and online, verbatim, with your name next to it, forever. And often out of the context in which you said it.

Now that can seem scary, but there are ways around the challenges that presents. The most boring is to just script what you're going to say, and repeat it over and over. But if you have a good relationship with a reporter, they'll cut you some slack and paraphrase your "um"s and "uh"s and false starts into coherent sentences. It's actually very rare for a reporter to try to do a "gotcha" and use your words against you, but it's not impossible, and if you're talking about something sensitive, be aware of that possibility.

You can also think about going off the record for information that's particularly sensitive, or for announcements that you'd like to embargo until a certain date and time. However, not all reporters are willing to do this, not all reporters who say they're willing to go off the record will honor it, and even those who say they'll honor an embargo will often "forget" and slip up, leaking your news early. Basically, these are advanced techniques and I rarely have to relate off the record information to reporters. Don't play with this stuff too much if you're just starting out, and be sure to say the exact words "This is off the record." clearly and out loud before you even start talking about anything that's not for publication.

The best interview ever

My best interview experience by far was with Austin Bunn, who wrote a cover story in the now-defunct Brill's Content. After the magazine flew me out to NYC, they picked me up early in the morning and we went to shoot the cover. Afterwards, I met Austin and we had a long lunch and hung out for another hour or two afterwards. He was writing a multi-thousand word piece so he had resources and time to do it right. He must have done hours of research on me and we did a couple phone interviews before my NYC visit, so the visit involved getting all the details right about my story. I guess it was one of those once-in-a-lifetime things where a reporter has all the time and energy in the world to properly research a story, but there was nothing I didn't like about the experience.

The worst interview ever

Imagine having to wake up at 5am. You throw on some sweats, drive ten minutes to your office and await a phone call. You have no idea what the interview will be about. The phone rings and an overly caffeinated host starts throwing three-part questions about the theory of meme propagation to you and asks your opinion. You go blank a couple times live on the air in a major media market. It sucks beyond belief. That was me about five years ago. It was a morning radio show in New York, and they required me to have a landline for sound quality, which forced me to borrow an office phone way too early in the morning on the West Coast.

Some final thoughts

  • If you have time, answer reporters as soon as you can and talk to them. I am probably quoted once a week somewhere in the press because I always call back.
  • Getting interviewed is fun and often leads to weird old friends and family contacting you out of the woodwork when they run across you in the newspaper.
  • Simply put, live radio sucks and you should avoid it like the plague unless you're a great extemporaneous public speaker. I've never done live TV before, but I bet it sucks even more. Podcasts are a great alternative and are easy because they're mostly edited to make you sound better.

(note: special thanks to Anil Dash for helping out on this)