Congratulations! The Segment tools are amazing, and the team is 100% dedicated to success.
For example: I wrote a complex iOS app, and was having app store issues related to metrics; the CEO of Segment came to my office to help, and we looked at the app together, including source code and libraries, to diagnose the issue. This kind of excellent help is why I recommend Segment to all my clients.
If I run a website and I actually want to send all my data and as much of my users' data as I can possibly collect to, say, ten different vendors, then Segment looks great.
On the other hand, that means that I'm giving ten different vendors the ability to monetize all the data my site generates essentially for free. This is great for them, kind of crappy for the users who get tracked and monetized, and terrible for me, because I'm not realizing the value I generate.
I'm frankly rather surprised that Condé Nast doesn't try to prevent third-party tracking of its readers so that it can realize more of that value itself. After all, why should an advertiser buy a high-value targeted add for display on a Condé Nast property when it can buy the same ad with the same Condé Nast-dervied target for much less on a site full of cat pictures?
>"On the other hand, that means that I'm giving ten different vendors the ability to monetize all the data my site generates essentially for free. This is great for them, kind of crappy for the users who get tracked and monetized, and terrible for me, because I'm not realizing the value I generate."
Presumably you are not directly sending data to any vendors you are not receiving value from. In the analytics world, it is kind of understood that if you use Google Analytics, you should have an expectation that Google is using that data however they see fit. Same holds true with other vendors unless they explicitly say otherwise.
As such, the burden (rightfully) falls on the site owner to properly vet their vendors to make sure they work with companies whose business practices align with the site owner's data policies.
Nobody is forcing you to integrate with vendors who don't adhere to your standards, but not everyone has the same standards.
So at the end of the day, I'm confused by your point about how this is terrible for you since you're not realizing any value that you generate. You are exchanging data, money, or both, in exchange for a service. It is up to you to determine what that is worth to you and whether you are paying a fair price, but it isn't Segment's fault if you decide to partner with companies that give you a raw deal.
When I said that sites don't realize the value they generate, I didn't mean that they don't get something else out of the deal. Google Analytics is a genuinely useful service, for example.
I suspect that a lot of sites aren't thinking hard about the tradeoffs, though. The benefits of using these services are immediate and obvious, whereas the costs are rather more subtle.
I heard a fascinating talk last year by John Gruber about ways of monetizing content outside the usual sell-tracking-rights-and-accept-ad-bids model. (Check out what he does at http://daringfireball.net/feeds/sponsors/. The whole site seems to be entirely devoid of third-party tracking.)
This is a shameless plug, but Heap doesn't monetize data you send us. We don't share it with third parties, or do anything with data between different accounts.
I wish there were some sort of public list of companies that operate this way, it's a bummer that the default assumption (which is very reasonable) is that companies are using your data for additional revenue via other sources.
Fan of Segment. Their team are super helpful. Recommend it. I've not really found a reason yet to not like them. Sure sometimes you need a lit bit of ad-hoc work, but the benefits outweigh the costs in my opinion. Sometimes the engineers aren't in charge of the decision making process on analytics, so selling management an "insurance" policy, which is how I see Segment, is very powerful to get things done. Service provider not working out for you? Switch it. No code changes.
Also, I don't believe the argument you can get away with one provider. MixPanel told me they don't do emojis in email subjects because their CEO doesn't approve of it. Don't want to get anyone in trouble, but I kid you not I have that email and instantly made me look elsewhere for email service - why should 1 person dictate my companies marketing strategy? Segment prevented a lot re-work for this.
Anyway, in short, Segment is good for the market because it means services have to up their game as lock-in isn't as strong...except for Segment's!
The system is great. Only reason I have had to dislike it is that they would forward messages out of order. I think that might be fixed now after I told them to look into lamport clocks. It caused some problems with forwarding to Intercom.
I really wanted to like the product, they are a great team and the potential sounds amazing. But for me, I found that while they do cover the basics in each product they integrate with, often I want to use special features that they either don't offer or that it becomes to cumbersome to use. I am talking mainly mobile SDKs where each service offer something else like Mixpanel's a/b testing and intercom in app messages etc.
They cover a whole spectrum of products so different in their value prop that I think just can't capsulated with a single api
We should talk. mParticle covers off on full feature support for our integrations and works with some the largest apps in the world. Purpose built for mobile. Full feature support for integrations. Better controls around data.
Love the tool using it since beta. Its an obvious pain point brilliantly executed. Between segment and google tag manager the analytics instrumentation has become exceedingly easy and better documented in the last 3 years.
And I know from experience lock in is immense, once you add an analytics tool there is no turning back. With segment you can play the field more easily, ofcourse while being locked in to segment :)
> Initially, the team was building a competitor to Google Analytics or KISSmetrics, but they had also built a library called Analytics.js that wrapped all the analytic services and APIs together.
They released an open source version of this library on GitHub, and it just started growing by itself. That forced them to reconsider their focus. What people actually wanted, as it turned out, was a simple analytics API.
I'm not sure I follow this description of the company's pivot. What is the difference between their own analytics service and an abstraction over another company's service? Why do media companies need to use so many similar analytics services, e.g. the 30 odd trackers on many news sites?
> Why do media companies need to use so many similar analytics services, e.g. the 30 odd trackers on many news sites?
Let's see what we can find on CNN right now:
Clicktale, analytics for their design team with things like click heat maps and scroll maps.
Optimizely, A/B testing to compare the results of layout changes by serving variations of each page to portions of their traffic.
Turner corporation's internal analytics tools, for who knows what.
A tag for PostRelease, which distributes media for content-based advertisers, like advertorials and videos.
Usabilla, a tool for surveying user satisfaction.
Dynamicyield, which personalizes page content to specific visitors.
Krux Digital, which is about tracking people across multiple devices for a better measure of how many real people are on the site, and then segmenting that data for the advertising sales team.
Moat, which measures viewability of ads, i.e. what portion are actually seen by people rather than just being loaded below the fold.
Scorecard Research, which is audience measurement and surveys.
Chartbeat, which is real-time analytics specifically for the editorial team. It shows them what stories are hot and trending at that specific moment, which directs what headlines get pushed up, what journalists get assigned to cover, what stories get pushed off the front page, etc.
Outbrain, which adds the "more stories you might be interested in" type stuff to the end of each article, which are actually ads.
The rest are tags for serving banner ads through Google, their own ad server, and various exchanges.
On that last point I would have to guess because not all analytics and metrics vendors provide the same data, and/or because different teams or divisions use different platforms but ultimately the website is where all their requirements converge?
We use Segment and love it. It keeps our code simple and makes it easy for us to turn on and off services (especially for non-technical team members), their debugger makes QA'ing instrumentation really efficient, and their team is amazing and really helpful.
What creates the requirement to send data to multiple analytics services? Wouldn't you want to send it to one capture and analytics service and then funnel that into whatever line-of-business applications you need?
meh. to me the offering of segment carries very little benefit. it's great being able to swap in and out of different SaaS but honestly, I did not find the value offering to be significant because you end up fiddling with less than 3 or 4, 10 at extreme cases.
Ultimately, I ended up going with Mixpanel because they offered pretty much everything I wanted in analytics. I have integrated Segment everywhere but ended up ripping it all out once I realized that I had bought into the marketing.
For example: I wrote a complex iOS app, and was having app store issues related to metrics; the CEO of Segment came to my office to help, and we looked at the app together, including source code and libraries, to diagnose the issue. This kind of excellent help is why I recommend Segment to all my clients.