Attribution & lead scoring · South Africa
Know which marketing actually sells. Call the buyers who are ready.
Google, Meta and your website all claim the same sale. We follow each customer from their first click to the signed deal, give every channel its fair share of the credit, and score every lead on how likely they are to buy. You spend where it pays and your sales team calls the right people first.
Why your reports disagree
Every ad platform counts the sales it can see and takes full credit for them. Add the platforms together and you have sold more cars, policies or units than you actually did. Meanwhile the leads that never touch your website, like portal enquiries, walk-ins and phone calls, do not show up in any of them.
- Last click gets everything. Branded search and “direct” look like heroes because they come last, after the ads that did the real work.
- Clicks are not customers. The cheapest click is often the person who leaves after one page.
- The sale lives in the CRM. Your analytics stops at the enquiry, so nobody can say which campaign produced a buyer.
- Leads are called in the order they arrive, not in the order they are ready to buy.
What you get
Attribution: what each channel is really worth
- First-party tracking that links each person’s visits, on any channel, to the enquiry they finally send.
- Sales from your CRM matched back to those people, by email first and then by cell number, with checks so a shared number never gives credit to the wrong person.
- Multi-touch credit: every visit that helped gets its share, not just the last one.
- Portals, walk-ins and phone leads added from the CRM, so the channels your website never sees still count.
- Cost per enquiry and cost per sale by channel and campaign, plus how many visits from each one are actually engaged.
- A monthly dashboard and a plain-English read: what to keep, what to cut, what to test.
Predictive scoring: who is ready to buy
- A model trained on your own past buyers, not a generic points system someone made up.
- Every contact rescored each day from what they do: pages, return visits, enquiries, email clicks.
- A journey stage for each person, from early research to ready to buy.
- A hot-leads list for your sales team, so the first call goes to the person most likely to buy.
- Follow-up timed to behaviour: email, and WhatsApp for people who opted in on your website.
- Checked against real sales every month and retrained when it starts to drift.
How it works
- Check. We look at your tracking, forms, cookie consent, CRM export and ad spend, and tell you what can be measured today and what is missing.
- Connect. We set up first-party tracking and Mautic, the open-source marketing automation platform, on a server you control, then match your CRM to it every month.
- Model. We train the scoring model on your sales history and test it on months it has never seen, so you know how well it works before anyone relies on it.
- Run. Monthly reports, a live dashboard, briefings for your media agencies and a model health check, with the numbers explained in plain English.
Case study
A multi-brand South African dealer group
The group ran Google, Meta and TikTok through separate agencies, took leads from its own website and from the motoring portals, and sold through dealerships across the country. Every report told a different story.
- 9 times out of 10 the model scores a real buyer above a non-buyer, tested on sales it was not trained on.
- 45% of online leads came from the portals straight into the CRM, invisible to website analytics. They are now channels in the report like any other.
- Less than a tenth of the cost per engaged visit for the catalogue ads compared with the group’s broad social campaign, where most visitors left after a page or two.
- A hot-leads list for the sales teams, refreshed every day from the score.
We don’t name our clients. The client’s name and the full results are available under a non-disclosure agreement.
Who it is for
Businesses with a sales team and a buying decision that takes more than one visit: motor dealers, property, financial services, education and B2B. You will need a CRM or a sales record we can export, at least a few months of sales history, and enquiries coming in through your website.
Built for POPIA
- Your customer data stays first-party, on a server you control.
- Tracking starts only after a visitor accepts cookies.
- No personal data goes to Google, Meta or anyone else unless the person agreed and you approve it.
- WhatsApp messages go only to people who opted in on your website.
- Scores decide who gets called first. They are never used to refuse anyone service.
FAQ
Questions we get asked
What is marketing attribution?
It is working out how much credit each channel deserves for a sale. Most buyers see several ads and visit several times before they enquire. Attribution shares the credit across those visits so you can compare channels fairly and move budget to the ones that produce buyers.
Is this not what Google Analytics does?
Google Analytics sees website visits and stops at the enquiry. It cannot see the sale in your CRM, the lead that came from a portal or the customer who phoned. We join those up, so you see cost per sale rather than cost per click.
What is predictive lead scoring?
A model looks at what your past buyers did before they bought, then scores every current lead on how closely they follow the same path. The higher the score, the sooner sales should call.
How do we know the score works?
We test it on months of sales the model never saw and report how often it ranks a buyer above a non-buyer. We repeat that check every month against new sales and retrain the model if it slips.
Do you replace our media agency?
No. We give your agencies better numbers to work with and brief them on what the data shows. They keep running the campaigns.
What does it cost?
It depends on how many channels, sites and branches you have and on the state of your CRM. The attribution check is free, and we quote once we know what has to be connected.
Find out what your marketing really sells
Send us your channels and your CRM. We will tell you what can be measured today, what is missing, and what the first month of reports would show. No obligation.