DigitalGhost Interview

Dean Bloomfield, AKA DigitalGhost has agreed to answer a few of my questions and as ever comes back with some interesting points. And I am even going to claim that this was the real reason he came out of retirement 🙂

So once again read on…..


Retirement clearly didn’t agree with you, as it looks like about six months and you gave up! So what is on the cards for DigitalGhost in 2006 and beyond?

Work, work and more work. Seriously, with a start-up launch, (Actually Marketing) client demands, research, testing and application development I see 2006 as a year of challenges. I also think that in 2006 search engines and SEOs will work more closely together than ever before.

The industry has become more sophisticated and the level of sophistication will continue to increase in the future. As a result of that sophistication, the relationship between search engines and SEOs will become less adversarial because the goals become increasingly similar. So for the future I see more research into IR and search technologies than reverse engineering algos. Just makes much more sense to anticipate than to react.

You are known for your loquacious, very eloquent, but loquacious posting on forums. With that in mind, are we going to see a return of DG on the boards and what subjects are going to get your juices flowing?

I’d be filling fora with vowel movements now if I had more time. Well, time and the inclination to repeat myself every time someone new came along with a question that has been asked and answered many times. I think the reason many of the ‘old-timers’ don’t post as frequently as they once did is simple weariness. It’s all good though; the new generation of SEOs can serve some time answering questions.

There are several subjects that excite me. Digital rights, which has been mainstream news for almost all of ’05, social media and marketing, search technologies, education in the Information Age, why all the hype about Web 2.0, etc.

Algo’s and language seem to be a keen interest of yours so what evidence have you seen, if any, that the engines may be adopting techniques like LSI/hilltop/semantics etc?

If ‘java’ is returned as a result for a query on ‘coffee’, semantics are involved. The only question is to what extent are semantics incorporated into the algorithm? I don’t think any of the search engine reps will be mentioning LSI as it is patented by Telcordia. Are they using something similar? I certainly think so.

What was the last search related technical paper / patent that you read and do you care to share your thoughts on it?

Google’s Dupe Content patent.

My first thought is that it remains remarkably easy to smudge ‘fingerprints’ to the extent that Google’s dupe filter is duped. In reality, I think Google manages to find a lot of smudged and partial prints, which are of little use. As for SEOs and SEMs that want to duplicate content and use it, well, they learned to wear gloves and they don’t leave any prints.

You’ll know when Google perfects their dupe content filter when you can no longer find your stolen content in their results.

If you had to write a simple recipe for an SEO success what would the main ingredients be?

A. Forced me into a cliché huh? Start with good content. Then acquire quality links. Yes, it is that easy. And no, it is not that easy. Rather than a recipe, how about some qualities?

All the successful SEOs are curious. They exhibit perseverance. They are competitive. They’re confident. They aren’t afraid to fail. They hate to fail. They don’t mind being the underdog. They constantly educate themselves. But most importantly, they try a lot of recipes. A simple recipe is fine, for that simple success. But there are thousands of ways to succeed, several ways for each site, so why limit that recipe book to a single, simple recipe? Instead, cook something new every day. Throw out the recipes that leave a bad taste in your mouth.

So what did you manage to achieve with that Google Mini of yours. Although it would not have the latest algo on it I reckon it could be used on large-scale auto generated stuff to see what work and a whole host of other things, so what did you find out?

That little experiment was intended solely to find out what Google thought about word relationships. Is job related to employment? Is Job related to work or trials? Everyone keeps saying, ‘think outside the box’, well, I wanted to know what the box was thinking.

How do you rate the current search providers and what could they do better?

I think they’ve done a fantastic job of making information easy to find and a mediocre job of rating it. Until an algo is created that can differentiate between popular and important, we’re left with SERPs that display popularity contest winners.

What could they do better? Well for starters, they could solve the ‘AND Circuit” query problem. Go ahead, search for AND Circuit, with quotes and without, and let me know how useful the results are.

You have a solid and longstanding background I search, but six months is a long time. How long will it take for you to get up to a standard you are happy with and do you think it is possible for people new to search to get going quickly on a new site for instance?

I stopped working in the industry for six months, but I didn’t stop researching or reading industry news for six months. As for a standard I’m happy with? That doesn’t exist, there’s no room for complacency in this industry.

Is it possible for someone new to search to get going quickly? Sure. Will they rank well? I don’t know. Depends on the individual. The basics are well agreed upon, so an inquisitive business owner could start with the proper foundation and avoid a lot of mistakes. In competitive arenas though you need more than a couple of weeks of trial an error. I still find new ‘Link and a Promise’ sites in the SERPs though and they typically remain there until they hit someone’s radar. Getting to the top isn’t a measure of success though; it’s getting back to the top after you’ve fallen and managing to stay there that counts.

Without giving us the findings (unless you want to) can you tell us what sort of tests you are conducting right now to reverse engineer search engine algo’s and would you consider yourself a algo chaser?

A. I’ve never been a reverse engineer. I’ve always worked with prediction modeling. I want to be able to predict, within reason, where a site will place in the SERPs on various engines given a known set of variables. It’s much more valuable to have your own algo than it is to tear someone else’s apart.

Building your own search engine forces you to anticipate the same problems the search engines deal with. Issues like RLA, (Rapid Link Acquisition) are much more noticeable when you have your own set of patterns to work with than they are when you’re forced to borrow someone else’s patterns.

You may not know, but in the UK we had a marketing campaign for a mobile phone company that was based around whom you would have a one to one with. So dead or alive, who would you like to have a one to one with and why?

Nikola Tesla. I’ve always been fascinated by his inventions, his intellect, and his quirks. And I’d really like to know the truth about his ambient light with no apparent source…


Dean, thanks for your time and your insight, some interesting thought provoking ideas you have there and I wish you lots of luck and reward with your new venture.

If you want to read more from DigitalGhost I would recommend firing up his Blog…. DigitalGhost’s Desk.

ukgimp out.

6 Comments on “DigitalGhost Interview

  1. Your feed link down in the bottom left is wonkadoodle I think if you take the “feed:” out it should be ok. Good interview, and no I have no idea what that costume in the corner is about … really …

  2. This is really a comment to DG and not to the interview, although it was a nice read 🙂

    IIRC, it’s just a “gate”, not a full “circuit”. A “gate” search yields somewhat better results, although still far from perfect.

    And… (no pun intended, I swear) in Danish “And” is the surname of Donald, who is called “Anders” here, because “and” means “duck”. Funny thing, words… so, there’s a much deeper point than what’s said above 🙂

  3. >>Funny thing, words…

    Too true, which is why the search engine business is so interesting to anyone interested in words.

  4. I have often referred to SEO as the ultimate word game.

    That said, I don’t see how long we can expect SEs to accomplish such grand feats as “AND circuit” or even “AND gate”. Even with magic, there is not enough context in the query. I don’t agree the general solution lies in personalization, although that may be a short-term path to better results sets.

    I don’t know why the Butler failed, but at least he managed to elicit some context from searchers. I suspect that someday someone will ask searchers to provide some context along with their query… it won’t take much, I think. Multi-part forms (a la Teoma) are too clumsy…

  5. I don’t think the “AND Circuit” solution is that difficult, or won’t be that difficult when search engines start paying more attention to word relationships. Adding more context to the query doesn’t help much with the AND problem.

    On a humorous note, I did get this SERP, returned.

    I don’t think personalization is the long-term answer either but I do think context and proximity should be more thoroughly explored.