Attention follows names rather than effects. Watch your own figures for changes and treat named updates as one possible explanation among several.

Two things happened last month

Google announced a conversational product, available in two countries, which produced a fortnight of coverage.

Google also ran a core update to its ranking systems, which began in mid-March and finished rolling out last week, and which produced very little.

The second is the one capable of changing what appears when somebody searches for your trade in your town.

Which is a reasonable illustration of how the attention in this area is distributed.

Why coverage does not follow effect

The fourth is the substantive reason. A core update has no stated target, so there is nothing specific to report beyond that it happened, and an article saying rankings may have moved somewhat for some sites is correctly judged unpublishable.

What a core update actually is

Worth understanding once, since these happen several times a year.

It is a broad adjustment to how pages are assessed generally, rather than a change aimed at a specific practice.

Nothing is being penalised, and sites that fall have not done anything wrong in the way that a spam action implies.

Some pages are reassessed as better matches than they were and some as worse, across every subject at once.

Which means there is frequently nothing to fix, and the recovery advice attached to them is largely the same advice that applies anyway.

The honest summary is that it happened, it may have moved you, and the response is to be a better answer rather than to change a setting.

Check your own figures instead

Which is the only thing that tells you whether any of this concerns you.

Look at your arrivals from search by week over the last two months.

A change coinciding with the update dates is evidence; a change at another time is something else.

Look page by page rather than at the site total, since a core update typically moves some pages and not others.

Ten minutes with that data answers the question that no amount of reading will.

And most of the time the answer is that nothing happened, which is worth knowing rather than assuming.

A worked example

A business read that an update had completed and asked what they should do about it.

Their weekly arrivals from search were flat across the entire period, on every page that mattered.

Nothing had happened to them, which took eight minutes to establish.

What their data did show was a steady decline on two pages since December, unconnected to any update, caused by a competitor publishing something better.

That was the real finding and it had been visible for three months without anybody looking.

The update they had read about was the least interesting thing in their own figures.

The changes nobody announces

Which are frequently the ones that affected you.

Results pages change shape continuously: more images, a different local block, more space given to advertising, a new panel answering the question directly.

Any of those can halve your clicks while your position stays identical, and none of them is announced.

The way you detect them is by watching impressions against clicks: steady impressions with falling clicks means the page around you changed.

That comparison is available in your search console and almost nobody looks at it.

It is a better use of ten minutes than reading about any named update.

Set a threshold before you look

Since watching data invites reacting to noise.

Decide what size of change over what period would prompt action: something like a fifth, sustained over three weeks, on a page that matters.

Below that, note it and do nothing.

That prevents the pattern where somebody checks daily, sees ordinary variation, and makes changes that cause the problem they were looking for.

Monthly checking with a stated threshold produces better decisions than daily checking without one.

Who the advice is written for

Worth remembering when reading any of it.

Most writing about search updates is aimed at sites with a hundred thousand pages and a team, where a small percentage matters enormously.

A local business with fourteen pages and forty visits a day is in a different situation, and most of the recommended responses do not apply.

Your position depends far more on reviews, on being genuinely nearby, and on having the answer somebody wanted than on any of this.

Which is why the same four things keep being the answer here.

Write down what you concluded

A habit that pays the next time one of these arrives.

After checking, note the date, what you looked at, and what you found, in two lines.

Six months later, when the next update is announced and somebody asks whether it affected you, you have a record of what your figures looked like before and after the last one.

It also stops the same ten minutes being spent repeatedly on the same question by whoever happens to read the news that week.

Keep it in the same place as the speed figures and the list of third-party tools, since it is the same kind of record.

The counter-case

Some updates genuinely matter.

The one last August had a stated target and a clear description, and businesses publishing thin pages for every town were right to pay attention to it.

Ignoring everything is as poor a strategy as reacting to everything.

And a business whose traffic did fall sharply on the update dates has a real question to answer.

Check your own weekly figures against the dates, compare impressions with clicks, and set a threshold before looking.

The ten minutes

  1. Open weekly arrivals for two months.
  2. Look page by page.
  3. Check against the update dates.
  4. Compare impressions with clicks.
  5. Note anything falling for months.
  6. Set a threshold for acting.
  7. Check monthly, not daily.

Step four is the one nobody does, and it detects the changes that were never announced, which are the ones most likely to have cost you something.

The general pattern is covered in the update that did not affect you.


Frequently asked questions

What happened last month?

A conversational product was announced, producing a fortnight of coverage, and a core update ran from mid-March and finished rolling out last week, producing very little.

Why does coverage not follow effect?

Products are announced and updates are logged, effects take weeks to appear, and nobody can say what a core update did, so there is nothing specific to report.

What is a core update?

A broad adjustment to how pages are assessed generally rather than a change aimed at a practice. Nothing is penalised, and there is frequently nothing to fix.

How do I know if it affected me?

Look at arrivals from search by week over two months, page by page. A change coinciding with the update dates is evidence; a change at another time is something else.

What changes are not announced?

The shape of the results page: more images, a different local block, more advertising, a panel answering directly. Any can halve clicks while your position is unchanged.

How do I detect those?

Compare impressions against clicks in your search console. Steady impressions with falling clicks means the page around you changed rather than your position.

West Coast Media Solutions Inc. provides web design, web development, hosting, digital marketing, and business consulting to organisations across Canada, drawing on more than twenty-five years in the field.

Read that an update completed?

Open your weekly arrivals from search for the last two months. Eight minutes tells you whether it concerns you.

Start a Conversation