Kinds of impression

Memorial weekend is a good thing. Actually, I think all long weekends are good things.

The product launch continued. We continue to get coverage a few days after the wire. Unfortunately, there was only one article written on the product. On the other hand, that article was picked up by quite a few outlets. So, between the wire piece and the article, we seem to have covered quite a bit of ground quantitatively.

One gratifying thing to note is that we are seeing increased traffic to the website. Before the launch, majority of traffic comes from search engine and the information requests tend to be the few that I have targeted with search engine in mind. But, this past week, there is a nice increase in information request coming from other areas that were not targeted. So, as a proxy, there is definitely increased interests. You might ask "shouldn't the site design be such that it drives the traffic to only the targeted/desired areas?" Well, if you have to ask that question, you do not know websites. Long story short, it is a perpetual catch-up game since you can only react to data which, by definition, is old news. Nevertheless, I am exacting some additional budget to streamline the website. Still trying to put my arms around this project but this should be a major undertaking for the next month or so.

Remember the "hot lead" that Harte Hanks called me for? We listened to the call recording a few days later where it was one of the calls. All that I can say is that the "hot-ness" came from the fact that the person asked the caller to call back at a specified time. (So far so good, objectively.) On the other hand, it was obvious that the person just wanted to get the caller off his back and, my guess, gave a call-back time knowing full well that he will not be there to pick up the phone.

Anyway, we decided to re-do the script. The individual elements are kept pretty much the same, but we put in a lot more navigational things in terms of what questions to ask. The original script was a sequential list where the caller just recite the questions from top to bottom. The new script has several branches and specifies that if the answer is Yes - go to Q5 and No - go to Q8. The idea is to take away the thinking for the caller. Two observations. One, it is unfortunate that we assumed a higher level of caller capacity than what was warranted. We always know that these are just hire guns and we cannot expect too much from HH. On the other hand, we have assumed that the caller would exercise some discretion in the basic zig-zag of a conversation. Did not happen. Even when the person on the other end already answered the question two minutes ago, the caller still recite the question. Anyway, I am glad to provide the hard-coded intelligence if it cannot be had organically. Two, it is kind of fun to come up with the branching script - it is like writing in unstructured programming languages like BASIC with a lot of GOTO. So, the new script was shot to HH for testing and I hope to hear the specifics before the week ends.

The $25 promo idea still sucks. I am quite disappointed. The latest incarnation is that the headline with no mentioning of the $ gets better performance. I will end the program this week and figure out what to do next. I also contracted a web designer to re-do a number of these landing pages. Hope to run the new pages in a couple of weeks.

I am sure that you heard me talking about Google paid-ad's before. The latest fine-tuning comes from separating out the Search Network from the Content Network. Basically, if you specify a campaign to do Search network only, it will pop up only when an user searches for the specified term. This is particularly desirable because it means that person is motivated to find the information. On the other hand, Google also pop up ad's next to relevant content using some proprietary algorithm - less desirable because the person who clicks on the ad may not have an immediate need. But, the good thing about the Content network is that it gives volume in terms of impression. A typical day, I may get a few hundreds of search impressions, but I typically expect to get a few 10,000's content impression if not more. Herein also is my problem. One of my campaigns starts to get a bit expensive and majority of the traffic comes from content network. Cost issues aside, my biggest concern was that the high Content network fee is crowding out my search network usage. So, while I am paying good money because of the content network volume, I am missing out on quality search network listing because of the budget. (This is a bit convoluted, if you really want to understand the thinking, drop me a line and I'll try to explain better.)

So, all these explanation leads to my decision to separate out the Content and Search network. Basically, I now run two campaigns for each category instead of one with identical content with the only difference in the network targeted. The early verdict is quite positive. The search network has been cranking along as expected and I no longer have to worry about people not being able to find us in a search. The content network is getting fewer impression because of the budget constraint. However, there is no noticeable drop in our website hit and, more importantly, product/sales demo requests. Finally, this gives me a much higher level of fine-tuning capability to trade-off between costs, number of impressions, and type of impressions. My plan is to run these campaigns for another week to collect enough data and change all the other campaigns into this duel-mode.

We are going to do a senior management on/off-site where all the managers fly into the HQ for Kumbaya and strategy sessions. I am in the process of preparing the marketing piece. This has been an interesting exercise because this is the first time since coming on board four months ago that I try to take a step back and see the big picture on what fits where. And, this also helps me to identify whatever is missing and where to put the resources. I should have something more concrete in a few days.

chiefchickenheadless (at) gmail (dot) com sign out