Robert_Charlton - 8:31 pm on Apr 30, 2013 (gmt 0)
Here, I am talking about all traffic means direct traffic, referral traffic and search traffic.
I think it's hard to get a precise picture with all three lumped together. You will probably also have to get "more granular" within each of these classifications, once you separate your views of them.
Several replies early on in this thread suggested that rogue bots were a possibility... and with the aggregate information only, that might well have explained it.
Till Mar'13 there were on an average 300-350 visits/day, 3.5K page views/day and 5-6 pages/visits per day.
Now visits are more than 450 per day, page views are 1.2k and 3 pages/visits per day.
We measure engagement onsite through "downloads of software" and our number of daily downloads were not affected after page views decrement.
I take this additional information, all together, perhaps as a positive sign, though again that's a rough interpretation only. The same number of goal events (downloads) with fewer overall page views per day suggests to me that visitors are seeing content that motivates them to make a decision more quickly, which perhaps is an increase in efficiency.
On the other hand, you're getting fewer downloads per visitor... but the change in engagement on the page view level appears to be more dramatic than the increase in traffic.
I don't know whether or not you've segmented this data further. I suspect in some cases you probably have. To look further at what the change of traffic is, you'd want at the least finer detail on search referrers, eg, and query data (if available). For purposes of this discussion, you'll want to identify what traffic is identified as Google (or might be coming from Google). I'm assuming that "increase in efficiency" is something that Google considers a good thing, which may affect what traffic Google is sending where. Is there a change in queries for which Google is sending you traffic? Is there a change in landing pages?
Within your site, are you tracking visitor behavior? Have you set up conversion funnels, and are you seeing any changes in visitor patterns since March 13? What about time on pages? Return visitors?
There are many things you can measure... probably more than might be useful for your situation... but, as you can see, it's hard to draw conclusions from just broad stats. Some of the visitor behavior data may well suggest useful site changes you can make.