Metrics To Track In Online Operations
This is not really proper post, it is a collection of points I have repeated to many people (last posted in a Hacker News thread). Figured it would be a good idea to have it put up somewhere.
What it is a list of metrics you need to track when you run an online operation and the importance of each element:
The metrics
1. Page views: Track overall growth
2. Page views per user: Track engagement
3. Visits: Track overall growth.
4. Unique visitors: largely useless metric since every tool out there disagrees with each other on the number.
5. Revenue per user: Track user segmentation.
6. Revenue per page view: Track infrastructure costs.
7. Active users: Finer measure of engagement and churn.
8. User acquisition/retention cost: Track cost of marketing/outreach.
9. Conversions: Free to registered, registered to paid.
10. Off-site: Notification emails, Newsletter subscribers, open rates, Twitter followers, Facebook fans.
Their importance
(1) Important for selling inventory on CPM.
(2) Needed to get the mix of organic growth with existing users to increase, while keeping a different focus on acquiring new users.
(3) Needed for the ad sales pitch. No point getting a million visitors if you can't retain them, but sales needs inventory.
(4) This has always been ~10% off between different tools. Can be an indicative measure of engagement.
(5) Helps keep ceilings in place for acceptable cost per user.
(6) Helps make decisions on infrastructure where you need more than what you can buy.
(7) Possibly the only user-related metric that should matter internally.
(8) Can help you understand why bidding on that $5 CPC keyword does not make sense when what you make per user is $3.
(9) Helps you fine tune monetization.
(10) Helps you cover all other external delivery platforms.
Read MoreAn Old Gem: The Opportunity For An Online News Operation In India
Strangely, most of the points are still quite valid and the opportunities have remained unaddressed.
Read MoreElgg, Buddypress, Drupal: Different Flavours for Different Reasons
One question that I repeatedly get to hear from users and clients trying to set up private social networks is the 'which' part of it. This post is meant to act as a very lightweight guide for that process. We will consider only Elgg, Buddypress and Drupal for this. There are numerous other options available like Dolphin, Ning and Jomsocial, but Ning falls out of the list due to it being a hosted solution and I have little experience with Ning and Dolphin in their current iterations.
The crucial question in selecting one of the many options is whether you know what you actually want? The effectiveness of the solution used can vary wildly depending on how clear you are about what you really want. Some of the solutions are pretty close to plug-and-play. Install them, fill in some data and you are good to go. The others are barebones frameworks, they need a lot of work to be made into a product. Clients also often make the grave error of underestimating the management overheads of running a full fledged social network. Between the technology, product and administration tasks it can easily wind up taking up a lot of your time (and often money too).
Social network frameworks also require much higher specification infrastructure to support it. While you can keep a blog going on a $3-per month shared hosting solution, most of these social networking products will start to become unusable on shared hosting as your user base and concurrent usage starts to climb. In such a scenario, it may be the best idea to go with a Ning or the commercial offering of Elgg, so that you are shielded from the technology parts of the puzzle.
A quick checklist of things that will make your choices easier would run like this:
1. How many users are you looking to support
2. Will it be only free users or paid users or a mix of paid and free?
3. How comfortable are you with managing technology (both coding and infrastructure)
4. Minimum budget
5. Time that you are willing to spend on it
6. What degree of support are you willing to provide users?
Most of the private social networks can be split into three kinds:
1. Solely for communication: You want only the ability for members to sign up and communicate with each other in common threads.
2. Full featured: You want a clone of the features of the bigger networks.
3. Full featured and heavily customized: You want a clone of the bigger networks and also a lot of custom features added on.
In the case of (1), you can easily make do with a Google Groups set up or something similar to that. Anything else is really an overhead that you get little value from. For (2), go with one of the commercial offerings and you will get the best of both worlds. (3) is where most of my work has been and in and thus the longer elaboration on the point.
In using a customized solution you are making a substantial commitment to a platform from which switching won't be easy or cheap. Each platform has its own strengths and weaknesses, but one thing that is common among all of them is that after a point in the growth curve you have to constantly reevaluate problems you had fixed earlier. A simple example is of a feature that lists all users in a single page. This may work flawlessly when you have hundreds of users, but it will get your site to a crawl when you move into the thousands.
The three solutions under the scanner here – Elgg, Buddypress and Drupal – all take different approaches to producing an outcome largely owing to their respective lineages.
Elgg is a social framework that has a few modules in its out-of-the-box form which enables it to work as a bare-bones product.
Buddypress is based on WordPress and it is now a full fledged plugin on the WordPress platform which is a content publishing engine.
Drupal is a content management framework which has nothing social in it out-of-the-box. It takes a lot of work to get it to be a social networking site. You can read the guide to it here.
The three solutions can be further evaluated under the following heads:
Ease of use (users): Buddypress is the easiest of the three to use since the base framework it uses is WordPress. Elgg comes in second, while Drupal is the hardest to figure out.
Ease of use (site administrators): Buddypress is the easiest again due to the WordPress lineage. Drupal comes second becomes of its superior update notifications and built-in tools to atomically manage users and permissions. Elgg finishes third because of its really clunky administration interface.
Development: Elgg comes in first here because of its origin as a social networking engine. It also has code that is targeted only at PHP5, which means that it carries little bloat from the PHP4 days. It also has the best API of the lot and a well executed views system. Buddypress comes in second because of the simplicity of the WordPress API, but it can quickly become a major limitation if your requirements start of exceed what is provided out-of-the-box by it. Drupal suffers because most of the heavy lifting for the 'socialization' of the platform depends on various modules. Customizing these can often be quite complicated.
Platform Maturity: At the core Elgg is the youngest of the three, but it comes up on top again. It is a very well designed and executed platform at its core and its entities system is built to scale well if you can throw the right amount of hardware at it. Drupal comes in second because of the solid core it has been built on top of. Buddypress comes in third because of the core WordPress engine, which is still only a publishing engine and not really a framework.
Flexibility: Drupal comes right on top here because of the dazzling array of modules that it can leverage. There is a module for almost everything you can think of in Drupal. While it is a different matter that it can be a nightmare to integrate them all in a logical manner, the fact is that Drupal does give you the option to pull it off. Elgg comes in second with its well-designed API that allows you to extend it quite well and without too much trouble. Buddypress will be last here because it is tough to get it to do anything that is not there at the core.
It is most tempting to tally these scores and proclaim a winner based on my experiences, but I won't do that since every social network has requirements that are unique to itself. This is meant more as a guide to help you along the way of making better informed decisions regarding your platform choices.
Read MoreDeath Of The Homepage Revisited
Looks like I am not the only one who is bothered by thoughts that surround the decline of the homepage. Gawker's Nick Denton posts a pretty long one on the upcoming changes to the Gawker properties, of which the most significant change is the end of the home page as we have known it at least on his network of sites. It is worth reading the post in detail and Felix Salmon's even longer take on it (it is good till the half way point, after which it disintegrates).
A quick summary of the proposed changes yields the following interesting points:
1. Plateauing of ad revenue and rising cost of content: Selling ads at quantity (CPM) is always a game with a predictable (also undesirable) end. Things get even dicier when you are a publisher of quality content (unlike a Facebook or Twitter, where the users do it for you) as the cost of producing content in the vein of Lifehacker or Gawker is not something that gets cheaper as you add more hands that create it. There is also the 80:20 problem, that 80% of your traffic comes from 20% of your content.
2. Contrary to what he says, it is the death of the front page: If every page was a front page, you'd see only excerpts of stories there. There is no more a front page, every page is a story page. I am reasonably certain that there is a graph inside Nick's computer that says engagement is greater and bounce is lower on the story pages than on the section and home pages. There may actually be a lot of method to this madness.
3. Day parting is a bold step. I think this will be a limited measured run. If he cracks it, Nick would have made the first foray for any content publisher online into this domain. As an added bonus he gets a solid legacy to lean on, which is much better than the slimy bits that is associated with him and Gawker.
4. Content cross-pollination – This is old school and done-to-death. But I think the attempt is to make it one continous system (in format) across the system. Milk more page views. Nothing new here.
5. Video: Differentiated inventory again, with more day parting. I like this.
6. The last point makes the clear that this is the final move in a long term strategy:
- Build the audience
- Build a differentiated set up for ads (cut out the networks)
- Build a different ad framework (new formats)
- Alter the brand's (Gawker's) persona
- Hit growth targets
- Cash it in
Are Internet Portals In India A Case Of Death By A Million Cuts?
A curious aside from yesterday's post on In.com is that almost anyone who has a portal in India is not having a good time. There are not many who can claim to running one in India these days. I can only think of Rediff (Rediff.com), Times Internet Limited (Indiatimes.com), Web18 (In.com) and Sify Technologies Ltd (Sify.com) as the only players out there. The common thread between all these companies is that they are all in the red. It is a flaky to point out the portal angle as the only connection between these companies and their losses, but it is not without merit.
My reasonable guesstimate puts the top line for a company (it is a guesstimate, so do pardon me if the numbers are way off) close to Rs. 100 crore in revenue for a year. Most of these companies have a head count that is not lower than 150. With such a head count, compensation itself can easily be over Rs. 20 crore for the companies in a year. Servicing the facilities will add up another Rs. 5 crore a year to the bill. The technical infrastructure can add another Rs. 6 crore to the annual bill. Add another Rs. 15 crore in marketing expenses and it all adds up to a good Rs. 46 crore in a year.
The problem is that the head count is way over 150 in these companies (especially in those companies that are doing Rs. 100 crore in revenue) and the expenses I have listed are on the lower side. I have also not added other significant components like content licensing costs as they vary wildly from company to company. In the portal side of the business, advertising on content is more or less the only avenue for revenue and with the cost structures these companies have, it is impossible to scale those without a corresponding increase in costs. After a point, it is a case of chasing your own tail unless you can alter the fundamental variables in the equation.
The numbers mentioned above are not authoritative by any means and a lot of it can't be corroborated in any meaningful or public manner. But it demonstrates the point that doing more of the same that has been done so far will take you nowhere. This also represents a reasonable opportunity for a well-funded new operation that can be small and nimble. This also explains why the successful larger companies on the Internet India have been mostly transactional nature, which is only of limited relief since none of those have managed so far to diversify out of their core offerings.
What do you think?
Read MoreRewind: Revisiting In.com
It is never easy to write about your former employers or products that you have been closely associated with, without biases or other aspects that cloud your judgement sneaking into it. But I will do my best to stick what is important and leave out the lesser details in trying to analyze the product.
The product itself was meant to be the spark that should have fired up Web18's growth engine and pushed the company into a higher orbit. It was the vehicle that would have taken the company to the NASDAQ listing, doing a CNN-IBN in the online domain for Network18. But, it did not start that way. In its early days, In.com was meant to be a much smaller product that would provide contextual and activity-based content to those who used the service. It was to have grown organically, mostly by leveraging network effects and it was to be used more due to the incremental value the service provided.
But the 2006-2008 period was, well, much like what it is like now in 2010 – a lot of money chasing ideas that could potentially act as incredible multipliers. In.com was also required to form a better narrative. At that point in time, Web18 had an industry-leader in moneycontrol.com that brought in most of the revenue, with ibnlive.com and a clutch of smaller sites bringing in a smaller percentage to the table. It was an operation that relied heavily on content and content-based companies rarely transform themselves into stories that the markets like to invest into. Thus the transformation of the simple idea into a massive consumer-oriented 'portal' was kicked off.
And it almost worked. We can never be sure if it could have been pulled off if other factors had not intervened, but the buzz certainly was there. After the CNN-IBN story, the market expected another bit of magic in Web18, the trade press was positive about it and even the i-bankers were talking about a valuation of a billion dollars. If you noticed the timeframe mentioned earlier, you would have guessed that what stopped the train right in its tracks was the downturn in 2008. To be fair, the downturn only brought out what were clear flaws in what we had put out as a product. With the supply of money being not as plentiful as 2007 and the markets going into shock, neither a listing was possible nor more accelerated burn could be sustained.
Currently, the product is something that is already bogged down by its legacy. It is still a significant part of Web18's listing strategy, but one that costs the company too much sustain. It is not an enviable position for the company – it has failed to diversify beyond what has been its strength for close to ten years (content) – and it has gone nowhere in building out the transactional side of the business other than having the bookmyshow.com connection on its side. To show potential, In.com has to be sustained, but the longer it stays without showing exceptional growth, the more feeble that story becomes for Web18. It bleeds, but there is no cure or respite in sight.
Sometime in 2009 I had privately said to a handful of people that Web18 should sell In.com. They still had a decent amount of good buzz going for it and with many investors looking to enter the Indian market, they could even have found a buyer who would have probably even turned in a small profit on it for them. Of course, that would have come at the price of compromising the listing and severe loss of face, but that is what would have taken to change course on the product and the company itself. This probably is not an option today as the underlying story too has changed significantly.
But going back in time, it is an instructive thing to take a look at what we got wrong with the product.
1. Cost: Since the market was flush with cash it also nicely slotted in with a 'name-it-buy-it' philosophy. We were quite gung-ho, with an attitude of taking on the best heads-on, cost be damned. At launch time the product was already on its back foot because of the huge costs involved in getting it up and running. This was a major mistake in a market like ours where the potential for growth is a significantly larger figure than the rate at which that growth is being realized. India is market where you don't want to fight with an arm and a leg tied. It is a long haul here and you can't last for too long on one leg.
2. The Product: We changed it way too many times and pulled way too many things into it. Email was one such mistake. It is a loss leader in majority of online operations. Most of the infrastructure that goes into sustaining it only receives and blocks spam and the inventory is of an awful quality. Music is a similar thing. The licenses cost money, it costs money to serve the traffic and the audience always will go away some place else if you try and make it paid. You are then left with skinning players and sections for less than 2-3 lakh a month.
We could eventually not put together a coherent, well thought out product. We put everything we thought would do massive traffic into a basket and put it together online. And it did traffic, lots of it, but at a cost that probably did not justify the returns.
3. Numbers: I don't think we ever putting a cost-per-user or revenue-per-user number ever on it. Looking back at it (with hindsight being the awful thing that it is), it seems pretty bizarre that we just did not attempt to do anything of that sort.
4. The why: The reason In.com was put together was to garner more traffic and bolster the numbers. It looked like the perfect means to an end. We really did not much research into whether there was the need for an aggregator or a music portal. While traffic is a very valid and good reason, it is really not the horse that draws the cart in the online domain. To succeed, you have to be a platform (which means you own a sticky audience) or you facilitate transactions (money or information) or you publish or distribute excellent content. In.com wound up being a bit of all of that and none of it in its entirety.
As it is the case with looking at anything in retrospect, it is much easier to say these things now. On the other hand, I am not really sure if things would have turned out differently even if we were willing to address these issues at that time. When you are riding a wave, the world really does feel like your playground. But it has a lot of lessons for us who still are in the domain, trying to build out different things in it.
Read More