Before trusting last year's numbers, confirm the tracking code is on every page, that the business controls the account, and that filters and goals still match the site. Then compare last year with the year before.

The login nobody remembers

For many small businesses, Google Analytics was set up when the website launched. Someone added the tracking code, showed the owner a chart, and that was the last time anyone looked.

The first week of January is when the question comes back: how did the website do last year?

The answer is probably in there, but only if the account was counting properly all year.

So the order matters. Check the counting first, then read the numbers.

What to check before reading anything

The second causes the most trouble later. An account set up under a former employee's or an agency's login can be lost along with the person, and every year of history with it.

Who controls the account

Analytics accounts have no single owner. Each person with access is given permissions at the account, property or view level, ranging from read-only access up to managing other users.

Open the admin section and look at the users at the account level. A Google account belonging to the business, rather than to one employee, should be able to manage users and edit settings.

If the only full administrator is a former web designer, an agency or someone who has left, ask them now to add the business's account. It takes minutes now, and much longer once they stop answering email.

The code on every page

Pages built from a different template, such as a page builder's blank layout, often skip the header where the code lives. So do pages outside the main site, such as a booking system or a shop on another address.

The quick test is the real-time report. Open it in one window, then visit your own pages on a phone, one at a time, and watch each appear. A page that never shows up is not being counted.

Watch for the opposite problem too. A bounce rate close to zero usually means the code runs twice on each page, which doubles the pageviews and makes every visit look engaged.

Filters that have gone stale

Filters change data before it reaches the reports, and their effect on past data cannot be undone.

The common one excludes the office's own visits by its internet address. If the office has since changed provider, the address changed too, and staff visits have counted as customers ever since.

Accounts with no filters often show visits from fake referrers, spam that appears as traffic from sites nobody has heard of. Those visits never happened, but they inflate the totals.

Look at the filters on the view you use. Each one should still describe something true.

Goals that stopped counting

Goals are how Analytics counts what matters to the business: a form sent, a booking made, a call button tapped.

Most small business goals count visits to a thank-you page. If the site was rebuilt, or the form now shows its message on the same page, the page the goal waits for no longer exists.

The goal then reports zero, and the account suggests nobody has enquired for months.

Send a test enquiry and check that the goal counts it in the real-time conversions report. If it does not, fix the goal before trusting any of its numbers.

A worked example

A fencing contractor in the Fraser Valley had its site rebuilt last spring with a new theme. This week the owner opened Analytics for the first time in a year.

The graph showed visits falling to almost nothing in May and staying there. The owner's first thought was that the new site had failed.

It had not. The tracking code had been pasted into the old theme's header file, and the new theme never had it. People had visited all summer; Analytics had simply not been told.

The account also had a goal counting visits to a thank-you page that the new contact form no longer used.

Whoever maintains the site fixed both in an hour: the code added to the new theme properly, the goal pointed at the new form's confirmation, and a test enquiry to prove each.

The data from May to December is gone for good, which is the real cost of not looking.

What last year actually looked like

With the counting checked, the year can be read. Set the date range to last year and compare it with the year before.

Look at a few things only: visits by month, where they came from, which pages people arrived on, and how many came on phones.

Then the goals, month by month, if they were working. That is the closest the account comes to saying whether the site did its job.

Write the main figures down somewhere outside Analytics. A spreadsheet with a dozen numbers per year is more useful than a dashboard nobody opens.

Reading the figures honestly

Treat any month when the tracking or goals were broken as missing, not as a bad month. A gap in the data is not a fall in business.

Search keywords are mostly hidden in Analytics, shown as not provided. Search Console holds that information instead.

And set the numbers beside what the business actually saw. Fewer visits alongside the same number of enquiries can mean the site is attracting better-matched visitors, not failing.

The counter-case

For a small site with a handful of pages and one contact form, most of this takes twenty minutes: the code, the one goal, the yearly totals.

Sometimes the honest finding is that the history cannot be trusted. Then the useful history starts now: with working goals from this month, next January brings a fair comparison.

If control cannot be recovered, a new account set up properly under the business's own login is better than waiting. The old history is lost either way; the next year does not have to be.

For everyone: check the counting first, then read the year, and write down what you find.

Before reading the numbers

  1. Log in with the business's own Google account.
  2. Check who manages users at account level.
  3. Visit each main page while watching real-time.
  4. Send a test enquiry and confirm the goal.
  5. Review every filter on the main view.
  6. Compare last year with the year before.
  7. Write the figures down outside Analytics.

Step two is the one to do today. Every other problem in the account can be fixed by someone with full access, and if nobody at the business has it, none of them can.

Which goal deserves counting depends on what your website was supposed to do.

Frequently asked questions

Why check the tracking before reading reports?

Because pages, filters and goals break quietly. A missing code or a goal pointed at an old thank-you page makes a good year look like a bad one.

How do I know who controls our Analytics?

Open the admin section and look at users at the account level. A Google account belonging to the business should have permission to manage users and edit.

How can I test the tracking code?

Open the real-time report, then visit each main page on a phone. A page that never appears is not counted. A near-zero bounce rate suggests duplicate code.

Why are our goals showing zero?

Usually because the site or form changed and the thank-you page the goal looks for no longer exists. Send a test enquiry and check it registers.

What should I look at for last year?

Visits by month, where they came from, which pages people arrived on, how many used phones, and goal completions by month, compared with the year before.

What if a former designer controls the account?

Ask them now to give the business's own Google account full permissions. It takes them minutes, and is much harder once they stop answering email.

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.

Not sure your Analytics is still counting?

Give us read access for a day. We will tell you what is broken before you trust last year's figures.

Start a Conversation