dNeero.com Conversations Blog

Some Degree of Paid Seeding is a Must

Joe Reger, Jr. — Tue, 12 Aug 2008 23:24:00 GMT

From ClickZ: "Face it, even if you become an expert at organic seeding, the competition is tough. Very rarely is a branded viral video going to take off organically. Some degree of paid seeding is a must if you want to rise above the clutter."

Article: Money & Credibility

Joe Reger, Jr. — Fri, 25 Jul 2008 20:01:00 GMT

The latest article was posted a couple days ago and deals with the issues of money and credibility within blog-for-pay models.

Bug Fix to Codes Visibility on Share It Page

Joe Reger, Jr. — Mon, 21 Jul 2008 16:14:00 GMT

For those of you who were having trouble seeing embed codes after joining a conversation, we've made a fix and you should be able to see them now. Apologies for the bug and thanks to everybody who reported it.

No Complaints on dPolice So Far

Joe Reger, Jr. — Thu, 10 Jul 2008 01:52:00 GMT

So far so good. We've flagged about 50 people for putting up obviously junk content. And nobody's complained. Which means that we're calling their bluff and they're not contesting it. If we had flagged people and they had complained then our criteria for flagging could have been wrong. We're terrified of shutting down legitimate content or censoring anybody... that's not how we roll. But we will continue, as we have in the past, to increase the quality of the content in the system. Early indications are that the flagging system is getting people to stop junk postings. It's not that it was a huge problem. But the junk postings are what potential customers focus on... as they should... we're asking them to pay for people to engage their brand with their friends. If people aren't doing that then the model breaks down for everybody.

Change to Impression Scoring Coming Soon: Limit Per Time Period

Joe Reger, Jr. — Thu, 10 Jul 2008 01:07:00 GMT

There was a code launch today. For a few hours clicks on a person's answers acted differently on the front-end than they had in the past. (I've since changed the UI back to, more or less, the way it was before.) People complained that maybe they'd have to click into profiles to read conversations and that this would be a burden.

I do understand. But potential customers get what's up and they don't like it. It's clear that they don't want to pay for junk clicks. A small subset of users aren't reading each other's answers which is what an impression should count for. Instead, they're trading clicks and in a rapidfire manner are generating impressions for many users' conversations.

So I'll soon implement something that makes it so that you can click as much as you want but only one impression is recorded per sixty seconds per web visitor. When a click doesn't count I'll display a warning on screen to that effect. Maybe 60 seconds isn't the right time period... maybe it's 2 minutes... maybe it's 20 seconds. In this way we'll be able to argue that people are actually reading the answers and that clicking purely for impressions isn't happening.

We know that people are actually reading the answers because we email a lot of folks and we check web traffic stats. But we can't prove it clearly and the quick-clickers are very obvious about their activity on Facebook groups... which is what potential customers find very quickly. Thanks Search box.

Right now we can't prove people aren't quick-clicking and it's hurting our ability to sell. Companies are saying "well, I don't want to pay cash for people to quickly click and ignore." As much as we argue that this is a small percentage of users and that the holistic result is great exposure, the risk of junk clicks spooks them.

Most people use the system on blogs and in Facebook without trading clicks with their friends. Clicking-without-reading is hurting most users as they're lumped into a category of people essentially gaming the system. Clicks alone were never the intention... reading of the conversation and true involvement was. I thought that clicks was a good proxy but maybe not.

In the short term I didn't want to enforce it because the dollar amounts were low and we were young so I decided that it'd be best not to be a hardass about it. But maybe that was short-sighted.

Us getting sales is good for everybody. We're talking to companies about multi-dollar opportunities and series of conversations. We can only foot the bill personally for so much longer. I'm not naive enough to believe that anybody really cares right now whether we go bankrupt, but I do think that people would like to be part of something that grows, takes on new partners and starts to generate some real cash. Getting there means increasing the quality of the system as a whole... from the way impressions are counted to the ways that people can interact.

One big change today, apparently overlooked in the terror exhibited when a small group thought they'd actually have to, you know, read conversations, is that answers are now editable. Something that's been requested for many months but was herculean to implement.

Quality's going up. Changes to impression scoring coming soon.

Content Flagging and dPolice Launched

Joe Reger, Jr. — Tue, 08 Jul 2008 14:13:00 GMT

We've just launched the notion of content flagging. You won't get flagged for having a particular opinion. But you will get flagged for junk postings, hate postings (which we haven't seen), etc. The conversation igniter can flag content but nothing happens until dPolice (dNeero's Content Police) check it out. In grey areas we'll side with bloggers. The goal is to increase the quality of the content created in the conversations. Junk posters can no longer hide! Let us know what your experience is.

Ability to Edit Answers

Joe Reger, Jr. — Tue, 08 Jul 2008 14:12:00 GMT

By popular demand... drum roll please... the ability to edit answers! When you're reviewing your conversations you now have an Edit button. You know what to do. Expect some breakage. New functionality and all (and it's a pretty big feature, even if on the UI it seems small and obvious.)

Money in Research: Edgar Allen Poe and Game Theory

Joe Reger, Jr. — Wed, 25 Jun 2008 18:31:00 GMT

Money's used in research all the time. Most college psychology experiments are paid. As long as the money doesn't introduce bias then the results remain valid.

Notification: New Release and Balance Lag Time

Joe Reger, Jr. — Thu, 15 May 2008 10:01:00 GMT

Hi all. We know, we know. Some performance issues lately. Just so many of you joining conversations and posting them to your blogs. We love it, but to keep up with you we need to do some performance optimization.

To that end, we're launching some new optimizations today. Most is under the covers. The performance issues were actually caused by background processes that were taking too long to execute. For example, we calculate your balance and determine whether to pay you overnight. That's compute-intensive (even though it seems simple). It starts running around 3am and isn't ending until noon or so. It used to complete in a couple minutes.

So, we're optimizing it and other processes like it. The tradeoff is that we have to do some of that processing at other times. We think we've made some good compromises which will give you better site performance.

Expect things to look a little wonky for a few days as we work through the changes.

Also expect that your balance may not be reflected properly on screen. Be assured, your balance is in no way affected. The base tables are all there and your balance is preserved. What we're doing is storing a calculated/cached balance in a new table for performance. (We never use the new balance for actual financial calculations, of course... we look to the base tables.) It'll take a few hours for the system to populate the new tables with your current balance information. Until that happens you'll see invalid numbers as your balance.

Let it ride a couple days. If it's still wonky please contact us.

Apologies for the inconvenience (the performance issues lately and the balance update time period). We're working hard to keep up with you!

We'll probably launch the new code around 11:00am today.

Update: The code has been launched and now the balance updates are being calculated. We appreciate the patience.

Update: Joy. Looks like the new balance stuff triggered some conversations to close for a while. We'll get 'em going again.

Update: Balances should all be back to normal. Apologies for the inconvenience. Let us know if you see anything funky.

Article for Talentzoo.com: N-way Engagement Marketing

Joe Reger, Jr. — Tue, 13 May 2008 15:30:00 GMT

dNeero CEO Joe Reger, Jr.'s latest article for talentzoo.com is up.

Rankings Capability for Conversation Igniters

Joe Reger, Jr. — Sat, 03 May 2008 19:33:00 GMT

We recently launched a new capability for those who ignite conversations. It's called Rankings.

What are Rankings? Rankings allow you to choose certain questions as indicators of a quality that you want to track. For example, you could create an "Environmentally Friendly" Ranking and when people answer the question "Do you care about the environment?" with a "Yes" you assign them 50 points.

Identify Users Across Many Conversations: You can use Rankings to measure across many conversations. A conversation here generates a few points... a conversation there a few. In this way you get a sense of how a person responds over time to many different things. You're using your dialog to learn. And you're using the Ranking to quantify.

Build Panels with Rankings: As people score points you can skim those in the top 10% of Ranking score and put them into a Panel which allows you to ignite conversations that only they can see and take. In this way you're not only targeting but engaging your audience.

Surveys are Now Conversations

Joe Reger, Jr. — Sat, 03 May 2008 19:23:00 GMT

We just launched a big change. We looked at what we were doing. We listened to what our users were telling us. We watched how our users engaged the technology. We explored the effect that it had on their readers.

The result was this. We weren't surveying. We were starting conversations. What we previously called a survey was really a starting point. The questions of the survey framed what people discussed. And from there people ran with it.

So we went with it. From here forward we're in the conversation business.

To make the switch more complete we've also added a new feature. You'll notice it the next time you take a survey. In order to complete the survey you'll need to provide your own question. Then, anybody who comes to the survey from your blog or social network will need to answer that question in addition to the original ones.

In essence, this expands the conversation, helping what we now call our Conversation Igniters find out what's really on your mind.

We'll be publishing an article outlining some of the engagement marketing theory that backs this concept. For now we just wanted to put up a quick note so that you don't get confused by the new terminology.

Let us know what you think!

For Researchers: Filter Results Capability

Joe Reger, Jr. — Wed, 16 Apr 2008 20:14:00 GMT

It?s about time we got back to building some cool stuff for dNeero Researchers and Marketers! Today we launch a filtering capability that allows you to see survey results for a subset of your respondents. To see it in action, go to one of the surveys you?ve already launched and click on Results from the For Researchers tab? the results section with your private financial info. If you view Results from the public survey you won?t see the new capability.

Click on Filter Results and you?ll be presented with a bunch of demographic fields. Age, gender, etc. Choose who you want to see and then click Show Results. You can add a name to save your filter for future use. Your filters will work across surveys.

How?s this useful? Let?s say you want to survey anybody in the blogosphere but want to compare results between Males aged 24-29 and Females aged 24-29. All you need to do is create a filter for each group and compare the results. Using all of the possible combinations in the demographics fields you can really segment your data and see what?s happening.

As always, this is a starting point. We?d like to allow you to compare two (or more?) filter groups on the same page. And we?d like to add some deeper querying based on survey responses? for example, show me how everybody who answered ?Yes? to this question responded. We?ll get there.

This change touched some fairly core elements of the system. We?ve tested extensively internally. That said, please let us know if anything seems goofy over the next few days.

TalentZoo Article Up

Joe Reger, Jr. — Tue, 01 Apr 2008 03:37:00 GMT

Got an article published on TalentZoo.com. Many thanks to Colleen, Web Editor at TalentZoo.com. I'll be doing a column every few weeks... assuming they continue to accept my writings. They've got a bunch of great content for marketers and according to them: "Each installment of your column will be featured for one month on our homepage receiving more than 200,000 impressions and in our email newsletter to over 80,000 recipients." Excellent. I worked dNeero in there where it was relevant but wanted to make sure that I wrote something actually useful to folks.

Please Ignore the Blog Comments

Joe Reger, Jr. — Thu, 27 Mar 2008 14:22:00 GMT

Somebody is abusing the social nature of the web. He or she is posting lude comments to all of the dNeero Blog posts. Note that they're not affecting any dNeero users' blogs... just our own blog... the blog you're reading now. We'll likely have to shut down the comment system for a while to let things settle. And to the person doing this, you should be ashamed.

Referral Money Display

Joe Reger, Jr. — Thu, 27 Mar 2008 13:43:00 GMT

Found a bug in the referral system display. Basically, when you refer people to dNeero your account is tied to theirs and when they get paid dNeero pays you a percentage of what they got paid. It's important to note that we're not taking anything away from them.

This has been going on for a long time but the visibility of those transactions hasn't been clear. The small payments have been wrapped in to impression payments which are already rather small on many days for most bloggers/facebookers.

The net effect is that you didn't have visibility of your referral earnings. Having identified the issue we believe that you'll have visibility moving forward but please check us on this. Wait a few daily pay cycles and check your reseller earnings page. You should also see reseller program entries in your balance screen. Let us know what looks wonky. Apologies for the lack of visibility!

More on the Facebook Performance Issue

Joe Reger, Jr. — Thu, 13 Mar 2008 21:12:00 GMT

A couple weeks ago we started getting performance issue reports from Facebook users (thanks peeps!). So we dug in and scratched our heads. No obvious error messages or memory constraints.

A few days ago we installed a performance metric capability with page-level resolution. We expected that it'd tell us which pages were slow so that we could focus on optimizing them.

Bewilderingly, the page that people complain most about was being output by the system at a brisk sub-second clip. Odd, we thought. Facebook says that our servers need to respond to theirs in 8 seconds. Less than one second certainly fulfills their requirements by a long shot.

But the reports continued.

Here's what we think is happening. Facebook has to process pages before passing them on to Facebook users. This takes them a lot of time on bigger pages. Pages with a lot of html.

Once you understand this the fix is relatively simple. We took a long list of stuff and added paging capability to break it up into smaller chunks. Early reports are that it's working and people can get back to that page they were having trouble with.

This is part of the fun of writing an app that runs with/on another company's platform. We're very dependent (i.e. at the mercy of) on Facebook's servers. I don't believe they've documented the exact page requirements in a way that accurately reflects what our users are experiencing. I could be wrong. I'll peruse the app developer message boards a bit.

The worst thing we heard is that we lost a couple users over this issue. We can't blame them. But hate to hear it. Sorry peeps! We'll continue to improve. Let us know where we're dropping the ball (as you usually do) and we'll keep juggling!

Page Load Times Published

Joe Reger, Jr. — Tue, 11 Mar 2008 16:37:00 GMT

Page load times are critical for a web application. As a site scales you invariably find performance issues. We've solved many in the past and will solve many in the future. As you progress along this journey you find that you need increasing levels of resolution when it comes to your performance data.

To that end, we decided to add page-level performance metrics to the system. These metrics tell us on an hourly basis what the average page load time is for each page on the site. We've got tons of other metrics in place too, but most are server-based... memory, cpu load, etc.

Once we had this thing built we looked at it and said "you know, users may want to see this." So we made it available!

For many months now we've had a number on the bottom right of each page we display. That number tells you how long it took to load the page you're looking at. You'll notice that it's now a link. Click it and you'll be taken to a list of all pages and average page load times.

Enjoy. Clearly we'll be using this to find trouble pages and optimize them.

Marketers Interested in Incorporating Social Media into Campaigns

Joe Reger, Jr. — Sat, 08 Mar 2008 21:33:00 GMT

"Social computing/word-of-mouth marketing tops the list with 67 percent interested in incorporating it into their campaigns. The findings resonate across all marketing, not just direct marketing." More here.

Pageload Optimizing and Improvement

Joe Reger, Jr. — Fri, 07 Mar 2008 16:13:00 GMT

We continue to get reports of slow Facebook page loads. Thanks Kymm. Thanks Paul. Thanks Ingasol. You are appreciated and we do apologize for the slowness.

As we mentioned in the last blog post, we're experiencing increased load on the servers. This helps everybody. With more users we can sooner attract top-tier companies to pay for campaigns. And with that we can get funding so that we can afford the shiny new servers to keep things zipping along.

The main point we'd like to make is that we're just as, if not more, concerned about it as you are.

We've been optimizing things behind the scenes for the last few days. To do this we change some code, launch it and then quantify the improvement. Doing this has the unhappy side-effect of forcing us to restart the app servers each time we launch new code. It generally makes for 50sec - 2min of downtime, depending on which server the load balancer sends your browser to. Some of the slowness you've seen lately has been us trying to make things faster... I know, ironic.

Today we stripped out some functionality from the Facebook main page. There's a lot of stuff that we were doing with your friends and surveys. Who took what? Invite friends. Etc. Most of you told us that you ignored that stuff. So we took it out. Hopefully this will speed up the page loads a bit.

We're going to look into other key parts of the system to see what we can remove/optimize. As always, we'll try to keep you up to speed with the blog here.

Thanks for your participation. We'll get the speed back.