Tags

, , , , , , , , , , , , , ,

I wrote and posted the following article to Bubblews last year. It too has now gone …

Many of the folk who post here have also written articles for writing site Helium over the years. On 14 May 2014, Helium announced its closure effective 21 May when no more articles could be added. It will terminate on 15 December 2014, giving members plenty of time to remove their articles if they wish to and to ask questions and commiserate on the forum.

It’s been an interesting journey, with Helium starting as a question and answer site in 2006. It then became an article writing site but allowed very short articles as well as longer ones. People were posting dozens of “articles” a day. Minimum article length soon rose to 400 words.

By the time I joined in February 2008, Helium was a thriving community of thousands of writers from around the world. There were regular paying contests, upfront payments for articles, revenue share, a marketplace of article writing assignments, payments by publishers if they wanted to use your articles, a very busy forum, and good customer service. Helium rose rapidly to become one of the top 1000 internet sites in terms of global ranking as measured by number of page views (www.alexa.com). As much as $3.50 could be earned upfront for an article, plus ongoing revenue share. My top earning article made me about $380.

Helium was riding the crest of a wave and everyone seemed happy, until late 2010. There had been talk since early that year that Google wasn’t particularly happy with the quality of certain sites, or types of sites, and that it was going to look at this. Perhaps in anticipation of having its revenue reduced, Helium made some sudden major changes in early December 2010. It cut out automatic upfront payments and revamped its Marketplace program, dramatically cutting the amount of money it would pay out. There was uproar on the forum.

Two months later in late February 2011, Google introduced a new algorithm known as Panda, after the engineer who put it together. Its purpose was to push poor quality sites lower in search results and good quality sites higher. Helium was lumped in with other writing sites, or content farms as they were often called, and was pushed down in search results. This was reflected in the site’s steady fall in global ranking and revenue share on articles in following months.

Unrest continued on the forum as revenue share and contest prize money fell. Marketplace was revamped again due to numerous complaints by members and called the Assignment System. There was a feeling of negativity as money paid to writers from all sources continued to fall. Helium made ongoing efforts to improve the quality of articles by putting a number of rules in place and carefully monitoring their implementation. There were plenty of promises by Helium that things would improve.

As part of its move to digital, print company RR Donnelley (RRD) bought Helium in June 2011, paying a total of about $70 million which included an earlier stake in the company. This was the great thing that was to result in a lot of extra work assignments and greater payments for writers. Members waited in eager anticipation. And waited. And waited. Nothing much ever seemed to come out of the takeover. The extra work and money didn’t happen. Instead, revenue share and other payments kept falling as Google, through Panda, continued to put the screws on the writing sites.

Helium kept up its move to try and improve article quality. It deleted most articles under 400 words, edited and fact checked huge numbers of existing articles, and added detailed writing guidelines. One of the more contentious rules was a ban on the use of first person (i.e. I, we, my, our) and the forum was full of threads with people arguing back and forth the merit or otherwise of writing in first person and whether it should be allowed in Helium articles. Page views and revenue kept falling. For a long time, the site had become increasingly cumbersome and hard to navigate and was full of old information, errors, and broken bits. By the second half of 2013, Helium’s global ranking was around 6000-8000.

Meanwhile, a new site or set of sites was being developed in the background. We were encouraged to fix up our articles and come up with a “headline” for each one. After many delays and a period of about two years from its first announcement, 27 “microsites” finally replaced the old site in October 2013. But the new sites still seemed to be a work in progress.

Members weren’t happy with them from the start. They could be tricky to use. Each one was completely separate. You had to wait for your article to be checked and edited before it went live. Contests were gone as was Marketplace / Assignment System. The Helium name had all but disappeared, perhaps in the hope of getting a better deal from Google/Panda.

The new headlines had become the article titles, instead of complementing existing titles as Helium had stated earlier. Many of the headlines didn’t stand by themselves. For some of mine, I had basically used a string of keywords, and this was now the headline/title. Capitalization in the headlines was all over the place and most punctuation had fallen off, including any colons, dashes, question marks, quote marks, hyphens and some apostrophes. There were thousands of examples of words and abbreviations messed up because they weren’t in the program used to fix the headlines when articles were brought across to the new sites, e.g. ID became I’d, R&B became Rb. US state abbreviations were messed up, e.g. Indiana was In instead of IN, Maine – Me, Oregon – Or, Ohio – Oh.

Any changes to articles had to be checked by volunteer editors before going live. But you couldn’t just fix headlines (although some editors did allow this). You had to also add keywords, add or fix the abstract (a new feature), delete any first person, and add links if the article had none. Some editors worried about article length, which was supposed to be 500-2000 words (previously 400-1500). I think a lot of people, including myself, baulked at all this, so errors in the headlines remained. Some people had thousands of articles to plough through.

Members’ lack of enthusiasm for the new sites was reflected in the number of new articles, although numbers had been declining for several years. In the months before the launch of the microsites, about 200-250 new articles were published each day, a fifth of pre-Panda levels. After the launch, this fell to about 100 or a tenth of pre-Panda, meaning that an ever-increasing proportion of articles were old and many probably needed updating.

Because the microsites were new and there was no link to the old site, the stock of well over a million articles were regarded as new by the search engines and started off way down the list in search rankings. In other words, indexing started from scratch. Also, over time, members had posted many of their Helium articles elsewhere and these copies became the original and the Helium articles became the copy, pushing these articles much further down the search rankings and attracting far fewer page views and revenue.

Early in 2014, the microsites had poor global rankings ranging from about 130,000 to about 3 million, with most around 300,000 to 900,000. In the following months, rankings fell steadily. Also, bounce rates were very high, with most of the microsites around 80-90%, meaning a large majority of people only viewed one page before leaving. An article, “How to reduce your website bounce rate”, states: “As a rule of thumb, a 50 percent bounce rate is average. If you surpass 60 percent, you should be concerned. If you’re in excess of 80 percent, you’ve got a major problem.” (http://www.inc.com/guides/2011/01/how-to-reduce-your-website-bounce-rate.html). Users’ time on site for the microsites averaged 1 to 1.5 minutes, meaning most people hardly read one article before moving on.

On 6 March 2014, I commented on the Helium forum: “I looked at the beyondprose [one of the 27 microsites] home page. ‘Today’s articles’ number just three. But they are dated 22 January 2014, 16 January 2014 (but I suspect this is an updated date to an old article, based on its number; initially the old dates didn’t stay on the article), and 5 March 2014. ‘Latest Articles’ lists five but they all go back to 2008 and 2009, with update date of today and seem to be writers fixing up headlines and so on. The other main thing on the home page invites people to join, though Helium isn’t mentioned until you click to go to the Helium network page. Overall, this can’t impress the search engines (Panda looks at whole sites). I’d be linking the 27 microsites and call the whole thing Articles Galore or some better name (let’s have a contest to find best name) if Helium can’t or won’t use the H name. It’d be better than what we’ve got now. I don’t think people relate to the new sites. There’s no brand name; nothing tying it all together; and with names like Science360 and the rest, well, a lot of people might think they are little more than scraper sites plucking odd articles from here, there and everywhere, given the bad headlines, old dates, etc.” I further commented: “A real mess, and just about dead.”

On 14 May 2014, there was an email from Helium with subject: “Important news from Helium Publishing”, announcing the closure. No reason was given. On the same day on the forum, these reasons were given: “Changing market conditions and the proliferation of competitive publishing outlets and free blogging tools, as well as declining usage and revenue, were all contributing factors. After careful consideration it was determined that the existing business model cannot be sustained.”

In my view, the fundamental reason for the demise of Helium was the release of the Panda algorithm by Google in February 2011, although a post by a staff member a couple of months after this indicated that Helium had never made a profit. It was a high cost and labor-intensive business model with many staff (about 60 at one stage) and an army of volunteers. Helium’s “backbone”, its rating system, never did work properly and article ranking was basically a lottery as members rated in so many different ways. The rating system also favored newer articles as I discovered in my many correlation analyses of ranking and article submission date.

After Panda, it might not have mattered what Helium did or didn’t do. I think Google’s aim was to get rid of the content farms, which it has basically done. Revenue fell steadily as soon as Panda was launched and continued to fall steadily. Perhaps the only reason Helium lasted so long was RRD funding. But I think RRD was only ever interested in Content Source (http://www.heliumcontentsource.com/), which will continue. There seemed to be little or no support in 2011-2013 to help Helium fix up its site, which became increasingly difficult to navigate and was full of old information and errors.

The final straw for Helium was probably the microsites. I don’t think they were ever going to work. Members didn’t like them, finding them hard to navigate and use, and being disappointed with extremely low and still falling revenue share. Web users didn’t like them as evidenced by the high bounce rates and lack of use. Google didn’t like them either and they steadily sank in terms of page view numbers. Maybe in the end, RRD got fed up propping up this part of Helium and pulled the plug.

I must say that for all its faults, Helium did offer good customer service by staff and a large number of hard-working volunteers. It’s sad to see the place go.

Advertisements