Monthly Archives: February 2014
‘Because Brahe dedicated his whole life to gathering all that data, Kepler is now cemented into history.’
— Ryan Spraetz
“Because Brahe dedicated his whole life to gathering all that data, Kepler is now cemented into history,” Spraetz says, and this becomes an on-ramp to his startup, a 15-person company called Keen-IO. As Spraetz explains it, Keen aims to provide the world’s online businesses with ready access to the sort of detailed data so diligently gathered by Brahe, giving them the information they needs to make the big leap forward — to, as Spraetz puts it, “turn them into Keplers.”
It’s a highfalutin pitch — honed over several months during Keen’s residence in the startup incubator Tech Stars — but it’s also a captivating tale, and it taps into a sweeping trend in the tech world. Web giants like Google and Facebook and Twitter have achieved huge success in large part because of their ability to analyze the enormous amounts of data their online services generate — to see exactly how their businesses are operating at the lowest of levels — and now, many startups and open source projects aim to bring this “Big Data” know-how to the rest of the world.
At the same time, Keen is different. Some big data outfits offer massively complicated data analysis tools that run across hundreds of servers and require hard-core engineering talent. Others provide polished iPad apps that let you analyze data in simpler, and less powerful, ways. Keen aims to find a sweet spot, offering tools that are both simple and malleable, tools that let you readily use massive amounts of data in precisely the way you want to use it.
“We’re an alternative to building your own software,” says Kyle Wild, Keen’s CEO, who founded the company alongside Spraetz and another engineer named Daniel Kador, two close friends from his days at IMSA, a live-in Illinois high school known for breeding tech talent.
The trio launched the company out of Wild’s San Francisco home, but as the operation has grown — recently attracting $ 2.35 million in venture funding — Keen has moved into a communal startup space in the city’s South of Market district. Run by an operation called Heavybit Industries, this space is solely for startups that sell tools to the world’s software developers. It aims to help these startups create a new kind of software infrastructure that makes it all the easier for developers and businesses to build exactly what they want to build. Keen is the poster child of this new movement.
The Origin Story
Keen can traces its roots to Wild’s time as an engineer at the online games company FableLabs. One day, the data analysis guy left for another venture, and the data duties fell to Wild. He spent a few months building a central engine that let the company readily crunch all sorts of data, as opposed analyzing data in ad hoc ways each time a question came up.
As Wild tells it, this immediately boosted the efficiency of its gaming service. In order to use the site, gamers were required to take an online tutorial, and with his new data analysis engine, Wild says, the company soon determined that the length of the tutorial could directly effect the its bottom line. If the tutorial was expanded, fewer people would actually make it onto the site, he explains, but they would end up spending more money. “That’s something you can only learn with really deep analytics,” Wild says. “It’s stuff like that let us go toe-to-toe with Zynga using only a few people.”
The tool was so effective, Wild eventually quit his job to found Keen. Basically, Keen offers a set of application programming interfaces, or APIs, that let you build your own data analysis tools. You shuttle all your data onto Keen’s online service and then, through simple API calls to the service, your software can query that data, slicing and dicing it as needed.
That may sound complicated, but this is a tool for coders, not ordinary business folk. The aim is to keep things simple while still giving coders the flexibility to harness data as they see fit. “You can ask us questions with easy-to-understand, easy-to-construct, logical queries, and we’ll take care of all the hard stuff, like storing data on our servers, scaling the system up, making queries fast,” Kador explains. And, yes, coders can build slick “dashboards” that deliver results to the ordinary folk.
Building Blocks for the Future
You’ll hear a similar pitch from Google, which offers a data analysis service called Big Query, and Amazon, which offers something called Red Shift, but Keen wants to give you more control of your data. Edward Dowling, who runs a small startup called App.io that plugs into Keen for data analysis, says he was drawn to the tool because it could deal with millions upon millions of events at any given moment, but also because it could conform with his own way of doing things. “Other services follow their own forms and paradigms,” he says. “Keen does not.”
‘Other services follow their own forms and paradigms. Keen does not.’
— Edward Dowling
The larger point here is that App.io can analyze data in its own way without building a new engine from scratch. This is another trend across the web, one Heavybit is trying to harness with its communal startup space, one in which companies offer you internet services for piecing together your own online business. In technical speak, these services are APIs, but you can think of them as building blocks. Rather than erecting an entire online business from scratch, you can assemble the basic infrastructure from existing services. Amazon’s cloud provides the processing power. Keen analyzes the data. Imgix processes the images. Twilio offers the voice and text communications. And so on.
“You should only be building the part of a website that’s your core competency,” says Kador. “You should be outsourcing as much as you can.”
Five or six years ago, if you pieced together a new service with various APIs, you called it a mashup. Today, this is simply what you do when creating an online startup. And the practice will only become more prevalent in the years to come. Though it keeps one foot firmly planted in the 16th century, Keen is the future in more ways than one.
The Exchange pairs a non-profit organization with two web analysts — one a student wanting the hands-on training and the other a mentor with years of direct experience in the analytics field. Together, they work on projects with objectives aimed at improving the non-profit’s website performance and overall use of their analytics data.
Every non-profit aims to become more and more efficient, delivering maximum impact for the minimum amount of cost. Smart application of analytics will be a must to achieve this objective.”
Other organizations have gained value from Analysis Exchange projects by not only exposing ‘what happened’ on their site and what were the successes but more importantly identifying factors that led to successes on the site and how to make improvements. An example of some takeaways have been:
- What content visitors consumed and where they came from
- What social channels drove the most activity to the site as well as off the site
- Factors that lead to significant increases in visits
- Competitive benchmarks of success
- What factors led to declines in traffic and how to correct
Analysis Exchange projects are completely free and take less than a few hours for non-profits and mentors. Google Analytics is the standard analytics tool used for all projects. Its ease-of-use dramatically improves the non-profits ability to continue to use web analytics after projects are completed.
Cloze, a startup that helps users manage their messages on email and social networks, is launching a new tool today that helps teams circulate content through those channels. It’s called, appropriately enough, Circulate.it.
Co-founder and Chief Marketing Officer Alex Coté said that if you’re “an executive sales guy” or a member of a sales or marketing team, you probably feel like you should be sharing interesting content, or content that reflects well on your company, but the actual process of doing so can be “very painful,” particularly when you’re trying to get the rest of your team to share.
With Circulate.it, the Cloze team is trying to make things as simple as possible. There are browser extensions for sharing content in Chrome, Safari, and Firefox, as well as a bookmarklet for mobile, but the real improvement is in what happens after someone shares content with the rest of their team. Shared content can be sent out individually or aggregated in a daily newsletter (Coté said the format was “Medium-inspired”). The newsletter comes with a big “share” button, and assuming you’ve authorized Circulate.it to post to your Facebook, Twitter, or LinkedIn account, it only takes a couple of clicks for each person to share.
Behind the scenes, Circulate.it handles the scheduling of each post, so you don’t have to worry about spacing things out (you can also adjust the timing for things like embargoes). It also offers analytics data, so people can see whose sharing is the most effective.
Now you may be wondering if sharing an article on Facebook is all that hard, and I agree with you that it really isn’t. But again once you try to get large groups of people to share, things get complicated (another startup Addvocate has also tackled the problem) and Coté said you start dealing with people who “can barely spell Twitter,” and they certainly aren’t using more advanced tools like HootSuite.
Circulate.it offers a free plan for individuals. Its team plans start at $ 49 per month for 25 users.
Marketers and sales professionals want to know who’s visiting their site, what content the target audience is consuming and what converts site visitors to paying customers.
In a B2B environment — where long sales cycles and multiple stakeholders affect sales decision — “knowing who’s coming to your site” takes on another dimension.
Say you’re in charge of marketing an eLearning system, and your target market includes telecom, hi-tech/software companies and universities. Your sales cycle could span several months, and there are multiple personas/stakeholders who will evaluate your company and your product.
Some key personas include:
- Trainers, professors and teachers evaluating user experience and ease of uploading curricula and content
- Management/administrators evaluating your company, pricing, client testimonials, case studies, etc.
- IT assessing technical aspects of products, maintainability, your technical support processes, etc.
As a marketer, your job is to ensure your site addresses the needs of each stakeholder, while realizing that the interests/questions each group of stakeholders are likely to be different. It’s critical that the message and content (that you invested so much in creating) “sticks” with the unique personas in each market segment.
Easier said than done; measuring and optimizing all the above isn’t for the faint of heart.
B2B Measurement Framework
Let’s walk through a typical scenario and highlight key performance indicators (KPIs). The measurement framework our eLearning marketing manager has in mind includes (and yes, they follow GA’s ABC!):
- What percent of my traffic comes from industries I target
- Hi-tech/software companies
- Percentage increase or decrease in traffic from industries I’m not targeting
- Traffic volume and frequency from organizations our sales team targets offline
- Landing page stickiness by industry and organization
- What content is very popular
- What content is most shared
- All the above segmented by the three targeted industries
- Number of whitepaper downloads by industry and company
- Number of demo requests
- Sales follow-up call requests
- All the above segmented by the three targeted industries
Rich Firmographic Data in Google Analytics
Insightera’s firmographic data is organized by 1) deriving information from site visitors by identifying their ISP 2) determining that organization’s information, including location, industry (and soon company size and company revenue will also be available).
With easy-to-navigate firmographic readily available, analytics data takes on a new dimension; advertising dollars can be better targeted, and you have the ability to customize a visitor’s experience in several new ways.
Here’s a few examples of the rich and super cool data you have access to with Insightera, nicely integrated in the Google Analytics Reports (in Custom Variables):
1- Traffic Distribution by Industry
Within the GA interface you have a nice presentation your traffic by industry. Telecom seems to be strong (24.1% of traffic) in the report below, while Education could use some love from your marketing team.
2- Engagement By Industry
You can also report on your KPIs by industry (e.g. see how “Education” is the number 2 industry in the report below)
3- Traffic & Engagement By Organization
This report below shows the platform’s ability to take data segmentation a step further, and highlights specific organizations within the industry visiting the website (e.g. Yale University)
With firmographic data integrated into Google Analytics, it is possible to optimize paid campaigns such as Google AdWords, LinkedIn, banner ads, etc., and pinpoint how many companies from a specified list visited your site, which industries and what size companies visited the site. It provides the opportunity to then target paid campaigns to those visitors and channels, or increase efforts to reach untapped segments of a targeted audience.
Not a whole lot of considerations. Insightera makes it easy to plug and play. In your ‘Admin’ interface, select your Custom Variables slots for the ‘Industry’ and ‘Organization’ — and let the rich data flow. Double check that the selected custom variable slots are empty and that you’re not already using them for something else in your Google Analytics implementation.
Equipped with this new data, you can automate and personalize remarketing efforts and create targeted ads based on any given criteria. In the example above, the education-specific whitepaper can be presented to your higher-ed visitors, while hi-tech/software related content can be presented to your hi-tech/software visitors.
Insightera’s recommendation engine filters visitors by location and industry, content preferences and CRM data and digital behavior patterns. This process then predicts which content or channel works best for each visitor.
Increase the Value of Universal Analytics with more User Centricity
If you’re an early adopter of Universal Analytics or planning to migrate to Universal Analytics, Insightera will soon have you covered. The same method described above can be applied and firmographic data can be integrated into Custom Dimensions.
As advertising and remarketing efforts reach new levels of focus, site owners have the most relevant information to meet their needs thanks to account-based marketing. Combining the power of Google Analytics with the new scope of firmographic data allows a new level of Performance Analytics. This set of tools offers deeper analytic insights into who your potential customers are, what they do, where they come from and what they consume.
With all the discussion surrounding what’s not working quite as well in SEO – from SSL encryption, to panda, penguin and humming bird updates. I wanted to focus our attention to what’s really working for us and hence the SEOs Silver Linings Playbook!
I’d like to discuss a couple of advanced SEO tactics and analysis – that work.
Advanced SEO interlinking
Acquiring high authority links to your site and building up domain strength is great but how can you effectively leverage the strength of your internal pages to further boost pages that don’t rank as well. That’s where advanced SEO interlinking comes in to play.
Identify internal linking prospects
Export all the pages from the “Top Pages” tab – these are pages that have the highest page authority.
This method requires a lot more steps but it also provides a lot more data:
Extract all the backlinks for your root domain in to an excel file.
Use pivot tables to analyze top linked pages.
To go one step further, you could merge both Ahrefs and SEOMoz data to give one giant list of top pages on your site (don’t forget to remove the duplicates though!)
Analyze keywords with the content
So now we have pages that we’d want to link from. Next we want to identify keywords and landing pages that we’d want to link to from these top pages. Create a list of your high priority keywords that you’d want to see increase in ranks. SearchMetrics is a great tool to help you identify keywords that are currently not ranking in the top 10.
Lets assume we want to help drive up the link equity for some of these pages and keywords. We run these keywords as filters using Screaming Frog (another favorite!) to help identify if these keywords mentioned anywhere in our Top pages list.
Screaming Frog then magically spits out pages that have these keywords mentioned in them. We can then use these pages for interlining to our preferred keyword/landing pages. Quick Note: Make sure you don’t overdo the interlinking, use some longer tail variations of the targeted keywords.
Creating semantic sitemaps is extremely useful when it comes to getting a better understanding of your site and how search engines crawl/perceive it.
Instead of creating one giant sitemap with all the URLs, its preferable to break out the URLs in to categories that are similar to your website. This helps you analyze which section of the site is being better indexed as compared to other sections of the website.
If we take Allrecipies as an example – we’d want to create a sitemap that mirrors the sites taxonomy.
If the site were structured well, with clear folders that segment each of the categories – creating these XML sitemaps would be a lot easier. However, if its not then its more of a manual process. At the end, you’d want to be able to view each of the segments that have been created using Google Webmaster or Bing Webmaster tools to give you a definitive picture of how these sections are being crawled
Here’s an example of what a semantic sitemap would look like –
In the example above, we see that certain sitemaps (home-décor.xml) have a low indexation rate. This data helps us analyze further which categories and sections of our site might not be driving traffic due to poor indexation by Google.
Social Content Marketing
No SEO playbook is complete without a strategy around content marketing!
Social signals do help in increasing ranks. At AdLift we’ve done a number of tests on the impact of social links and SEO that hat proved that these help in driving up page authority and in turn rankings. This particular case study explains how social links increased ranks for keywords faster that keywords without the social links.
However, just as you need a solid content marketing strategy you also need a strong social marketing network to help drive that effort. Synergizing efforts between your content marketing and social media team is a great first start !
I hope this post was useful – If you have any questions, please let me know in the comments!
About the Author
An alum from Columbia University, Prashant Puri has over 10 years of digital marketing experience in building sites into multi-million dollar enterprises. Prashant Puri currently runs AdLift – a niche Bay Area SEO Company focused on delivering digital marketing ROI.
Enhanced campaigns, image extensions, third party reviews… the list goes on. Fantastic features that have improved performance for search marketers.
But we always want more, don’t we?
What do we hope that 2014 will bring from Google to make us really happy?
What We Want
The lists below aren’t sorted by importance or even feasibility. This is speculative stuff.
- More and better demographic data. Demographic data in search is a tricky business, but if anybody can solve this it’s Google. Their data quality has come a long way in the last couple of years but it still has the scope to get better.
- Third party data in search. If I have a cookied list of people who have made a phone call to my business, I want to be able to adjust my bids and targeting for those users. In fact, I’d like to be able to buy third party data and apply that to my search targeting. There’s a lot of it out there I can use on my display campaigns, and think how great the performance could be on search!
- Better RLSA remarketing lists (minimum volume, longer duration, YouTube audiences, etc.). I can see why this product launched with restrictions on these lists, but boy do I wish we had more flexibility. I’d love to have a list containing people who bought insurance from me 11-12 months ago. As soon as they search for insurance terms again I want to make sure I’m appearing. But with a 180 day duration limit I can’t do that at the moment. Any time Google want to open this up, I’ll be ecstatic. Let me use my Google Analytics lists too!
- Richer ad formats. Surprise me. I kind of don’t mind what’s included. Google have been pretty inventive about these in the past, and sitelink descriptions, image extensions etc have made massive impacts to my regular search ads. Combine what’s been done with PLAs and we’re in a good place now. All this improvement has just whetted my apetite. Give me more!
- Relaxed character restrictions. I know these limits have been fixed since time immemorial, but think how good an ad you could write with a few more characters in your headline, now that you’ve got years of experience writing such concise, neat ads!
- Video content. Google have been experimenting with videos in ads for a while, but it’s been quite limited. I have quite a nice video, so let me put it in my ad so people can watch it if they want to.
- Campaign and ad group IDs, and ad parameters in AdWords Editor. This one really would make a difference. We use the API for a lot, but for ad hoc tools a spreadsheet is still the easiest way for a campaign manager to make bulk changes. Unfortunately I can’t make bulk changes to things like ad group names, because then there is no way to upload that back into AdWords Editor with the tool totally aware which group has changed to which new name. Each campaign and ad group has an ID, let us export it and make changes around it, the way we can in the API.
The other API only tool that we like is the ability to change ad parameters. These sit at keyword level, and we want to be able to change these on the fly please, without having to build new API tools each time we need to do something different and inventive with them.
- Better filters in the Dimensions tab. If I’m looking at the Dimensions tab, I can’t filter by campaign or ad group. What? That seems like a ridiculous oversight. Sure I can look at just one of these at a time by using the left nav bar, but are my choices really to look at a single campaign or the entire account? Why can’t I, for instance, include every campaign that doesn’t contain the word “Brand” in the campaign name, thereby looking at all my non-brand activity together?
- Bulk add/remove in Client Center level reporting. This one is personal, folks. On a regular basis I need to extract data from across all our accounts. That’s 250+, of which some should be included and not others. My choices are to include all accounts, or to add them one at a time. Dammit! I want to be able to add all, but still have individual controls to add or remove.
What Might we Actually Get?
Of the above list, only some.
Expect the demographic data to improve, but I’d be surprised to see much change to RLSA remarketing lists. I’ll eat my hat if we get third party data in search in 2014.
RLSA remarketing lists are dominated by the implicit user terms people agree to every time they do a search on Google. They’re already stricter for users who have signed in (they’ve explicitly told Google what can be done with their data, and it’s hard to change that for new products). Third party data is a step too far, probably.
We’ll definitely get some richer ad formats, probably including video. I suspect new formats on mobile will be prevalent too. That’s just continuing an existing set of trends. Relaxed character limits seem unlikely. There is no pressure on Google to change this, and the amount of upheaval for a lot of AdWords accounts makes it tricky to implement.
Regarding the management/reporting changes: your guess is as good as mine. Speak to your AdWords reps until they consider these as problems that affect multiple people. As long as its just a few lone voices asking for these changes they’ll be considered low priority.
What Changes Would be Actively Bad?
There are always still a few of these, generally regarding changes to defaults or removal of useful granularity of control. Each of these makes campaign management more complicated in order to recreate the level of control we used to have.
Example: since enhanced campaigns removed the ability to ability to have different mobile bids easily for different keywords, some PPC commentators discuss using one keyword per ad group to regain that control. It’s the kind of change we shouldn’t have to make, but in some cases we simply do.
I’d like to see Google avoid those kinds of changes this year. A small core of AdWords users spend the most time using the platform, and changes to benefit the rest that harm the sophisticated users are sure to reduce the good will towards Google.
Third-party domains pose a problem
A cookieless approach to the rescue
“Thanks to Universal Analytics we can track the iframe on our merchants’ domains and be sure we get all traffic.”
– David Fock, Vice President Commerce, Klarna
In Klarna’s new cookieless approach, the “storage: none” option was selected in creating the account in Universal Analytics. The checkout iframe meanwhile uses a unique non-personally identifiable ‘client ID’. These measures cause Universal Analytics to disable cookies and instead use the client ID as a session identifier. Because no cookies are in use, browsers that don’t allow for third-party cookies aren’t an issue at all.
Virtual pageviews are sent on checkout form interactions. Custom dimensions and metrics are used for tagging a visit, with a dimension indicating which merchant is hosting the iframe, and a metric showing what cart value the user brings to the checkout.
Complete tracking and assured analysis
With Universal Analytics features, Klarna ensures iframe tracking is complete across all browsers. By using the virtual pageviews as URL goals and funnel steps, goal flow visualizations are used to find bottlenecks in the checkout flow. The new custom dimensions and metrics together with ecommerce tracking mean that reports can now be set up to reveal how each merchant’s cart value correlates to its final transaction value.
We have a couple of reports that tell us that internet advertising grew quite nicely in 2013.
First up, Techcrunch has details from a new Nielsen report showing display advertising spend is up 32%!
Up next, The Search Agency would like you to know that the amount spent on paid search grew by over 37%!
It wasn’t all good news, with Bing seeing a significant drop in impressions served: a 7% drop YoY and 29% from the previous quarter…
So pop the champagne….except for you Bing. You need to go to the back of the room and think about what you’ve done.
Twitter has just released an update to its Android client (coming soon to iOS) that brings new photo editing tools to the service, which are likely meant to make it easier to share photos direct and keep people out of competitive apps like Instagram. The second change adds a significant element of event discovery and real-time trend monitoring to user timelines.
The event surfacing is the more interesting element, since it marks a considerable attempt by Twitter’s to meddle with the straightforward chronological nature of that part of its service (besides promoted content). In case a user doesn’t have any new Tweets to load when you manually update it, it now brings up recommended posts from people you don’t follow, as well as trending topics and suggestions about new people to follow. In the U.S. only, it surfaces event updates for things unfolding on TV, in sports and on the news.
Each content update features a link to click through for more Tweets centered on that conversation. It’s an extension of some of the other work Twitter has been doing around surfacing events and breaking news, including the Eventparrot experiment and a feature that was tested back in August to highlight nearby events via proximity-based alerts.
A couple of things to flag about this change: It only happens when there’s no other new content for a user to view, and when they express a desire for more content, which is very clever; and it represents a way for Twitter to secure its place as the source of live, real-time information about things unfolding on the ground, a reputation with Facebook clearly covets.
Others are already capitalizing on Twitter’s ability to identify and follow events as they unfold, including Banjo, but Twitter adding this as a native feature in its mobile clients could change the nature of the service at a basic level. Should it roll out globally, and expand its scope, mobile users could be using Twitter a lot more for things like local discovery than they had been previously.
Targeting a relevant audience is essential for PPC ads. Whether in the Search or Display Network, considerable time and attention is taken to ensure you are reaching an appropriate audience. A while back, Google introduced a new method to help determine uninterested viewers. In July 2012, Google ushered in a new tool to mute certain ads on the Display Network, using a small [x] in the upper right hand corner. This option allowed users to specify ads they no longer wished to see. No information was obtained, though, on why they wanted to hide these ads.
To remedy this, Google recently announced plans to take this concept to the next level. Throughout the coming weeks, Google will be transitioning the mute ad setup to include a three-question survey. The insights gained from these surveys will allow Google to better understand why users are opting out of certain ads.
What Ads this Affects
The news release from the AdWords blog mentions some of the Display ads will have this feature. Looking at information for the original tool, it appears that this feature is for Display Network ads that utilize either remarketing, or interest categories.
Why this Update Matters
Allowing people to specify why they no longer want to see the ads will give Google some potentially powerful insight. Google allows users to choose one of three categories for muting the ad:
- I don’t like the content.
- I’ve seen the ad too often.
- Ad is covering content.
Here’s a preview of the survey interface:
There is a distinct difference between these options, so Google no longer has to assume you aren’t interested in the content of the ad. Perhaps there is a formatting issue on the site, or with the ad, and it is blocking part of the page. If it is from a remarketing list, it’s possible the user has just seen that ad too often, but would still be interested in ads for similar products/services. Previously muting this ad would have caused Google to assume the person wasn’t interested in that category, or product.
Not liking the content is a bit broad, so Google takes that selection to the next level. A second part of the survey comes up if they select that generic reason for muting the ads. The user can then choose from three options.
- The ad is too distracting.
- They aren’t interested in the offer.
- It is a possible violation of Google ads.
When reported correctly, the last option could help Google find possible ad violations and clean up the Display Network ads.
What Does this Mean for You?
So what does this mean for advertisers? It remains unclear to what extent the data will be utilized. What is clear, though, is that the feature can potentially help advertisers get impressions from a more relevant audience. If uninterested users mute your ads, it will keep you from wasting time/money in showing the ads to them in the future. The purpose of remarketing and interest categories is to get more of a qualified audience. Understanding why someone would prefer not to be shown certain ads will help Google get more appropriate ads in front of them.
I could see this new feature being beneficial to both Google and the advertisers. The more insight Google has, the better ad experience they can provide for the users, which can only help the advertisers in turn. The key to the tool’s success, though, relies on people taking the time to give honest (and true) feedback. Excessive reporting of Google violations when they don’t exist, or people not participating in the survey, will hinder the success of this new feature.
As Google continues working on this tool, it will be interesting to see what powerful insights can be gained from the feedback.