After nearly ten years, I generated the last run of data using DataPrizm. Last June, after one of our prominent clients did not renew, I finally listed to Motoko and decided to wind down the tool after the last contract expired. I notified all of the remaining clients, some of who had been users almost since the beginning, and after I made the last call, it did feel like a massive weight had been lifted from my shoulders. I slept well for the first night in a while. This feeling of relief was a big surprise to me as I had not been relieved how much this tool was weighing me down emotionally.
As I pivot into the next phase of my career, I plan to focus on my original passion of developing market entry strategies, we have been making a lot of changes. The biggest of these steps has been shutting down several tools we have developed and it has been an exciting experience with a lot of self-realization. I now realize that I should have shut DataPrizm down years ago but there was always some exciting data mining project to come along that would extend its life. As recently as yesterday after reading this thought-provoking article by Pete Blackshaw on understanding your “DataBase of Curiosity,”what is required to own voice search and how DataPrizm already does help manage this complex process.
To Pete’s point about understanding the consumer through a Database of Curiosity, we analyzed a year’s worth of site search and social media data for an amusement park. We uncovered over 600k questions people had asked them, with and estimated 15% of them being directly monetizable. Sadly, 60% of these questions did not return any result, let alone one that could lead to a conversion.
Unfortunately, rather than looking at this as an opportunity, the client team was overwhelmed and wanted to ignore the data. After Senior Executive mandates, they started creating content resulting in a 22% conversion rate and $6.8 million in revenue in just the first month. After this project we tried to get people to understand the gold mine in site search data and again tried when chatbots were all the rage and yet again to prioritize sequential queries for voice search but no one is really interested.
My inability to find alternative solutions for clients was the next emotion to hit. Even a few essential functions matching my workflow clients wanted to maintain were missing. One client evaluated 20+ tools and discussed buying DataPrizm since nothing on the market could do half of what it could do.
The frustration was followed by resentment of the broader search community for many of these features not being required components in tools. I often asked myself, how could others not think this way and need the features of DataPrizm? When I reached out to vendors asking if functions were possible, most responded that no one had asked for it or it should be possible in another tool. One provider told me in confidence that their main goal is to aggregate as much data as possible and throw it into pretty reports no matter if it added value since that is what people wanted to pay for.
This blend of frustration and resentment caused me to avoid the search community as everywhere I turned I heard nothing but gimmicks and shiny objects that did little actually to move the needle for a business. But what turns me off are all of the people pontificating on the death of the keyword phrase and the rise of Searcher Intent (why they are using the keyword phrase) and Topic Clusters (groups of keyword phrases related to a topic) yet no one offers ways to manage or monitor their performance. There are over 100 articles on how you must move to topic clusters, capturing People Also Ask, and many of the anointed talking heads of Search preach about Searcher Intent but do not give examples of how to identify, segment it, or even demonstrate business value. Ass of this was possible in DataPrizm using various language libraries but few thought that was important in real-life applications. As a result of this frustration, I stopped speaking at conferences and writing articles except PubCon.
The other application that is wrapping up is our Site Migration Tool. This is another one that I cannot understand why SEOs and Web Developers would not even try it. We looked at the root cause of migration and website integration failures, and most were due to incorrect redirects of high-value pages. By not transferring content and links to the new URL all of that accumulated value was lost. This tool helped identify high-value pages based on 20+ variables and monitored them to ensure they were migrated correctly. In nearly every case, the potential users did not want to invest the time necessary to make the migration successful, yet said their biggest goal, after a pretty new site, was zero declines in organic traffic. Ultimately we made significantly more money fixing failed migrations after the fact than we would have made preventing disasters. That is the sad truth of SEO – most executives only care about SEO performance when the traffic disappears.
Despite the numerous successes we achieved for clients, during the wind-down period, I had all the feelings you would expect related to the failure. I thought about things I could have or should have done differently. So, following the brilliant suggestion from a recent LinkedIn post by Wil Reynolds on how Entrepreneurs learn – that more is learned from mistakes and failures and how you recover than by all the warm fuzzy outcomes, that post made me change some of the things I started to write and capture some of the mistakes and what they taught me. In the spirit of helping, here are a few things that kept these from being widely adopted and commercial success.
Despite creating nearly 100 tools over the year, one of my biggest realizations was that I am not wired for software development. I have been told I can be very stubborn, a perfectionist, and a bit too laser-focused regarding processes. I did learn a lot about the software development process and am trying to turn these challenges into a strength to advise others on their software development programs. Â
It is very common in the search industry. You see features missing from tools, so you code your own. Especially many old-school SEOs tend to be programmers and can easily code tools for themselves and then make the mistake of sharing them sometimes generating interest in them. The challenge is converting these tools from functional tools that do what they are supposed to do to ones for paying customers takes on a life in itself. To get paying customers, you have to make a lot of concessions to get them to pay for the service, like changing colors or adding pretty functions that don’t move the needle.
One of my problems, despite a very agile development methodology, I needed things to work correctly, and I often delayed features since they could not accommodate all of the nuances that rumbled around in my head or the craziness of the various client sites and processes.
My biggest mistake was not listening to my wife, Motoko. She is my most trusted advisor and champion. Motoko is a realist and told me early on that most people will not want to do the work required to get the maximum benefit from DataPrizm. She continued to remind me that most people only care about solving their individual KPIs and not solving all the company’s problems. She told me people never want to see a dashboard full of problems they know they don’t have the resources to fix. She suggested that we not license the software but offer the output as a service and provide reports and detailed action items.
We pitched this “research service” to senior executives and to Search Managers and did a number of pilot projects with nearly all of them loving the data but did not want to pay for insights they did not have the budget, time, or support to implement. One prospect told me that if management saw this report, they would be fired for incompetence when he saw the paid search cannibalization rate and the gap analysis of underperforming critical words. We had this problem with a few of our first clients where agencies and the paid search team were fired due to their ineffectiveness.
Along the lines of not listening to Motoko, another mistake was not staying focused on the original goal of the application and focusing on consumer insights. It seemed the further away from pure data analysis we were by adding customer-requested functions, and those that matched enterprise search tools, the further we got away from our core audience.
Even with support from the C-Suite, amazement by business intelligence teams, and despite using all the right vocabulary, we always ran into a roadblock when we had to explain how it worked. We tried every way possible not to mention keywords or Google but you could see that was some of our imports. Once those words entered the discussion we were immediately pawned off to the SEO team, who viewed us as just another rank tool and we could never get back to the executive suite. For example, we were introduced to a major global bank that wanted to connect with businesses looking to go global. In the pitch, we showed the demand for words that would identify this segment and how by generating the right content, they could connect with them. The CMO loved it and blessed a pilot, and handed us off to Digital Marketing Manager As we were gathering data Junior SEO told his manager he could do this because it was “just keywords,” and that was the end of the project. The consultant that connected us later told me the internal project went nowhere since they could not actually aggregate the data and build the business case for content.
Another huge mistake was not taking investment money. Yes, we did get an initial grant from our first client, but the rest of the development was funded by our consulting service and high-margin projects that utilized the application. I believe had we taken an investment, that would have given us the runway to better market to senior leaders with actual take-no-prisoners salespeople like other enterprise tools use. It would have also forced us to make the tool a lot sexier regarding look reporting at the expense of function. Most importantly, It would have also forced us to end it much earlier when it did not grow fast enough.
Early on we talked to angel investors and various VCs, large and small, and despite being profitable, none would invest in it since it could not become a unicorn. In a couple of cases, the potential investors talked to Google, who told them we violated their terms of service and threatened to take away our API access which was one of our primary data sources.
Another big challenge was not pissing off Google. Usually, within a few days of winning a co-optimization project, we would get a call from someone at Google threatening to take away our API license or remove me from the Technology Council. Ironically, in every co-optimization project, despite identifying how poorly managed the paid campaigns were and identifying hundreds of thousands of dollars in waste, every single project increased paid spend to capture the missed opportunity we identified.
First, while neither tool was a huge commercial success, I have to remind myself the applications were very successful/effective, and I am very proud of them and the team that helped build them. They did more than pay for themselves, and because of them, we won a number of consulting projects and international clients we have had for many years.
We estimate that DataPrizm created over $1.5 billion in value for clients. Yes, that is Billion! One of our more significant projects identified nearly $400 million in new business opportunities, and our first commercial project identified $600k in wasted paid spend in the first week. The recommendations from the first quarter the client estimated a lifetime revenue value of over $290 million which is why they gave us a development grant to add features. They are the same ones that fired us a few months later when the board found out about how much money they had been wasting.
I am also incredibly proud of what we did for Absolut Vodka. We mined millions of phrases globally to create the Drinks Discovery Journey that helped us understand the rich opportunity of targeting the Drink Curious and pulling them into the conversion funnel. Over the past 6 years, this collaboration has resulted in 22 Search, Digital, and Data Mining awards, including, to my knowledge, the only Search Marketing Data case ever shortlisted for a Cannes Lion. I am incredibly thankful to the team for all of the experiments we did use data and linking it to the entire customer journey.