SEO statistics: Predicting traffic growth
Ian Lurie Sep 7 2011
Prospective clients always want to know: What kind of impact will SEO have? How much traffic will I get?
I’m an empirical guy. I want to give an answer. But based on what?
Well, I did some serious digging into traffic patterns over time across 200 Portent clients, and I have some answers. Warning: This data heads into seriously touchy-feely statistics. Use with caution.
Oh, and I refuse to make an infographic.
Where the data came from
I pulled data as follows:
- 75% from Google Analytics, via the API.
- 15% from historical reports
- 10% from other analytics tools, like WebTrends
For all data, I used the last day of the month as that month’s result. So month 1 traffic data is what the client saw after a full month of SEO.
With balanced SEO, the first month rocks
Clients working on both on- and off-site SEO saw the biggest percentage growth the first month—over 100% rise in non-branded organic traffic. But overall growth stayed strong and continued to rise over six months. Overall results? 300% growth.
‘Non-branded organic traffic’ means traffic from natural search results, not including queries on the clients’ brand names.
Geekery: Some big differences
When I went through the data, though, I noticed a lot of outliers:
- Clients who had huge jumps and drops in traffic that couldn’t be explained by seasonality.
- Clients who got amazing results. One saw a 2200% increase in two months.
If you’re a statistics geek, month 1 saw deviation of 2.4. Ouch. In non-nerd terms, that’s like aiming for New York and having a good chance of missing the city and the planet, and landing in the Sun by mistake.
After month 1, the deviation drops fast: .4 deviation in month 2, then .3 in month 3, then .12. For month 6, the statistical gods gave me the finger: Deviation shot back up to .24. Which, if I were aiming for New York, would put me in the Pacific Ocean. Probably somewhere shark-infested.
Still, none of this made me super-confident. My data seemed to show there was no pattern, at all.
Then I started pulling things apart a little, and it made more sense.
All onsite, no offsite SEO means fast, brief growth
Clients who let us do onsite SEO but skipped offsite for one reason or another showed big initial growth, then a fast drop off to nothing:
That doesn’t mean onsite SEO is only worth doing for two months, by the way. Stop doing onsite SEO, and problems creep back in, you drop out of the rankings, and traffic goes down again. That’s not a statistic—that’s Uncle Ian’s School of SEO Hard Knocks.
I’d love to do a chart of folks who stopped doing onsite, but most of them don’t want me to have their data. Can’t imagine why.
All offsite, no onsite: A long, slow slog
We’ve also had clients over the years who only let us do offsite SEO. The results, if we stick as close to white-hat link acquisition methods as humanly possible, aren’t that exciting:
Not shocking. If all you do is offsite SEO—guest blogging, link bait, link detective work and such—you pick up momentum, but it takes time. It also keeps paying off, by the way, forever.
Offsite’s a little difficult to attribute. Part of offsite is writing onsite content. That could, theoretically, impact onsite SEO as well. So I only picked sites that were utter onsite catastrophes, where I knew the relevance impact of new content would be zilch.
The balanced approach wins
When I analyzed the balanced data minus the just-offsite or just-onsite SEO crowd, the results stuck, and the deviation numbers were far lower. So I’m willing to stick with this version:
SEO pays off big over time. It may pay off big, fast, but that depends on your team, your site, and the resources available. I’ll revisit this data periodically and see how the numbers change. For now, I’ll be keeping this around as a general guideline that I follow with lots of disclaimers.
Please don’t go stake your career on these graphs. I have a BA in History and a JD. My qualifications as a mathematician are purely genetic. Plus, I have enough guilt in my life.
Side note: Why only 6 months?
I picked six months as the timeframe for three reasons:
- After six months, there’s great growth, but the numbers diverge more and more as some clients stick with it and others stop.
- Average client tenure at Portent is 3 years, but some clients are with us on a short-term consulting basis of six months or less. I could get the biggest dataset with a 6-month timeframe.
- I got tired.
Ian Lurie is CEO and founder of Portent Inc. He's recorded training for Lynda.com, writes regularly for the Portent Blog and has been published on AllThingsD, Forbes.com and TechCrunch. Ian speaks at conferences around the world, including SearchLove, MozCon, SIC and ad:Tech. Follow him on Twitter at portentint.He also just published a book about strategy for services businesses: One Trick Ponies Get Shot, available on Kindle. Read More