Convertro Tackling Cookie-Free Attribution For Digital Marketers Says CEO Zwelling

Jeff Zwelling is CEO and Co-Founder of Convertro, a marketing metrics attribution company.

Convertro What problem is Convertro trying to solve?

JZ: Marketers today are in large part “flying blind” when it comes to properly allocating their marketing dollars among their various marketing channels. I say this because most advertisers are relying upon outdated tracking technology that is cookie-based, relies upon sampled data, and attributes credit only to the last click preceding purchase. The last click fallacy results in overspending on navigational ad terms while under-spending or eliminating marketing spend on the so-called “top of the funnel” terms that actually introduce customers to the website. We have solved this problem for marketers by overcoming these hurdles and giving them unprecedented transparency into their marketing channels.

In your company’s description, you note that Convertro is self-funded. Does this mean you’ve been profitable from nearly the beginning? Are their any special challenges to being self-funded?

Convertro’s technology was built over several years prior to establishing Convertro as a standalone concern with live clients. We are supported by client-based revenue from our existing clients, and we’ve had investment from the founders and others. So no, funding has not been a challenge for us.

What’s your view on the effectiveness of display advertising? Is it being proven with your methodology?

Everyone at Convertro comes from a pretty nuts-and-bolts direct background. Traditionally, we’ve been strong believers in search and affiliate but skeptics when it comes to the value of display. The exception of course is behavioral marketing, whose effectiveness we have personally experienced across many campaigns. As a result, Convertro is developing features to support behavioral campaigns. One is a pixel container that enables Convertro to selectively fire pixels to third party retargeters against specific calculated criteria. This enables our clients to market specifically to visitors that meet specific criteria – a classic distinction would be a repeat customer versus a new one.

Another feature is Convertro’s ability to determine an expected revenue calculation from a particular visitor in real time based upon a specific sequence of marketing events experienced by that user. For instance, a visitor whose first visit is from a comparison shopping engine will have much lower expected revenue than a visitor whose first visit came from a non-branded keyword search followed some time later by a click on a comparison shopping engine. Convertro is able to set a maximum spend value per visitor that a buying platform such as Invitemedia can then optimize against. In the above example, we would recommend spending a lot more on marketing to that second customer than to the first. This per-visitor calculation is only possible because of our unique ability to track users reliably from their first visit combined with a very fast database backend that enables value to be immediately calculated and fired off.

What’s client momentum like today? And how does your pricing model work?

Convertro has been very fortunate that there are many forward-thinking companies that have become Convertro customers. Interest in the space is extremely high, and adoption to date has been tremendous. Convertro provides tracking on more than 30 major websites today and growing. The verticals within which we have had the most traction are e-commerce and subscription-based businesses. Our customers include Seven for All Mankind, Gamefly and Luxurylink. We charge based on the number of events tracked.

How does Convertro differentiate from others in the attribution space such as ClearSaleing, Atlas or in-house solutions at agencies, for example?

Convertro differs from others in the attribution space in three key areas:

First, the Convertro solution does not rely upon cookies, so we are able to account for all customer activity even when a cookie is deleted. For a marketer, cookie deletion presents a significant problem: when a cookie is deleted, all user activity prior to deletion is lost, resulting in conversions tied to partial click trails and inaccurate attribution. Being able to bridge this cookie chasm is of particular importance to companies that have long purchase consideration periods and correspondingly high average order values (AOV’s). Comscore recently reported that 31% of users have their cookie deleted every thirty days. Our internal data shows that before implementing Convertro, one out of every seven, or 14%, of our customer’s marketing events were “un-attributable” due to cookie deletion.

Second, unlike first party analytics providers such as Omniture and Google (analytics), Convertro does not rely upon sampled data . The reason we are able to track every click is because we have built our solution on top of next-generation databases, which gives us the horsepower to track at such a granular level.

Third, we don’t think that determining allocation formulas is as important as making sure we are giving credit to every marketing event that led to a conversion. So we’ve focused on data quality rather than data analysis. Inside Convertro’s dashboard, you can select the “lens” that you want to use to view the data, as well as the allocations to feed into keyword bid management software or display buying platforms. We make it easy to understand how every event contributes on a first/multi/last-click basis and enable credit allocation according to the client’s business needs. For example, in a clicktrail that includes both paid and non-paid marketing events, we can calculate a revenue distribution across only the paid sources for the purposes of buying more media, while calculating a distribution across all marketing events, including paid and non-paid, to determine the contribution of channels like SEO. Fundamentally, we believe that capturing all the marketing events that led to a conversion should be our top priority: for an analysis to be reliable it must be based on the best data possible. The regression analyses that some professional service providers such as ClearSealeing market as “sophisticated math” pale in comparison to the session reassembly algorithms that Convertro’s engine is built on.

Convertro is therefore able to produce drastically improved visibility into which sources contributed to conversion.

Everyone agrees last-click attribution is a broken way of measuring campaign effectiveness. The problem seems to be that this is the way all of the contracts are already set up (ex. affiliate, cash-back, loyalty programs, etc.) Even if you can prove that other sources in the chain provided value, how can a retailer (for example) redistribute the payments?

In the case of affiliates, you can bonus those who are actually bringing customers to the top of the funnel to compensate for the fact that they are having their cookies overwritten by more cannibalistic affiliates at the end of the consideration funnel. With Convertro, we can detect when an affiliate is violating program terms, for example through cookie stuffing or by buying branded terms, and we send notification emails with PIDs and CIDs to our clients to either reverse the transaction or remove that affiliate from the program.

With paid search, we find that there are plenty of converting keywords that meet a client’s ROI objectives that are not getting purchased because they aren’t usually the last click. For many of our clients with strong brands, identifying keywords other than their own brand requires a technical ability to look all the way back to the generic keywords that often brought the customer to the site initially; these clicks are seldom within days of the conversion or even from the same IP address or cookie.

For our subscription service clients, the challenges of using last click are even more complex. The conversion is really three events – completion of a lead form, marketing the user to get them to become a free trial and then further marketing them to become a paid customer. Convertro is able to allocate credit back to all of the marketing sources as that customer becomes progressively more valuable. While our subscription clients will continue to pay bounties to lead gen for a form complete, they all spend significantly on other types of marketing to get that customer through each of those stages. Convertro understands this and divvies up credit in ways that enable our clients to properly allocate credit where credit is due.

Conversion attribution, especially cross-channel, is a huge challenge. Isn’t this best left to the big boys – Google, MSFT?

Our team is Google-quality and quite better than those at Microsoft/Atlas. At Google and Amazon, small teams that can share a single pizza develop most products–we build like that. Our approach to distribution is quite simple: we build a great product that actually works and let our clients refer us more business. We’re not into smoke and mirrors or super expensive professional services; we just make stuff that works.

What misconceptions do marketers have today regarding attribution and ROI analysis?

As you mentioned, savvy marketers know that the last click model is not the best way to measure the effectiveness of a campaign, and if they are measuring the ROI of a campaign based upon this flawed model, they won’t be allocating their marketing investment in the most profitable manner. We believe that the best way to measure the effectiveness of an advertiser’s marketing efforts is by determining the proportional revenue per visit, or pRev, for each traffic source; look at the revenue that PPC, CSE’s, and Social contribute drive for example. Determining an accurate pRev for each of these sources is only possible if you are using a conversion tracking solution, such as Convertro’s, that can overcome the cookie-deletion and data sampling problems.

Where does the creative and messaging fit in with your attribution analysis?

It’s an integral part of Convertro’s attribution platform. A marketer will only be able to truly understand the effectiveness of their creative and messaging by being able to accurately attribute and measure the success of those campaigns to which the creative and messaging are attached. In other words, you won’t truly know the effectiveness of the creative for your PPC efforts, via A/B testing for example, unless you can track each iteration of your ad on a multi-attribution basis. Convertro’s conversion attribution provides incredibly valuable insights into all marketing channels, including these.

Follow (@adexchanger) on Twitter.

Enjoying this content?

Sign up to be an AdExchanger Member today and get unlimited access to articles like this, plus proprietary data and research, conference discounts, on-demand access to event content, and more!

Join Today!


  1. J. McFarland

    This is one of the best interviews I’ve read on this site. It really shows when a company has used its own product before bringing it to market.

    Two thoughts:

    1) No cookies? What about view-thru conversions/attribution (i.e. the revenue lifeblood of most of this site’s readership)?

    2) Ecommerce marketers seem to me to care more about the _incrementality_ of a channel/vendor than they do simple attribution. Yes, I can redistribute spend more appropriately within my existing channels, but how am I going to measure _new_ channels of customer acquisition?

  2. @J. McFarland –

    Thanks for the compliment.

    1) With respect to view-thru conversions, we do in fact track them but DR is truly our bread and butter. We can upload view-through files and then treat them as a unique marketing events or use some A/B testing to determine what fraction of an allocation should be. But end of the day, we prefer to treat view-thru as a vehicle to drive clicks.

    2) That’s easy with Convertro. We produce a value per click. You test a new channel and we’ll tell you how much proportional revenue that channel drove. If it meets your ROI thresholds – spend more; if it doesn’t, optimize or remove.

  3. Hi Jeff, I’m running a panel next week on last click attribution at OMMA in SF. I’d love to exchange a couple of emails with you about your approach and how it differs from those of my panelists. I’m impressed by your article and the technology you allude to, and if I have enough information on this, will incorporate it into the discussion.

    If you’re interested in informing me a bit further, please email.
    Marc Meyer